Professional Documents
Culture Documents
Fine Coal: Challenges in
Fine Coal: Challenges in
FINE
COAL
Processing, Dewatering,
and Disposal
Mark S. Klima
Barbara J. Arnold
Peter J. Bethell
Published by the
Society for Mining, Metallurgy, and Exploration, Inc.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Society for Mining, Metallurgy, and Exploration, Inc. (SME)
12999 East Adam Aircraft Circle
Englewood, Colorado 80112
(303) 948-4200 / (800) 763-3132
www.smenet.org
SME advances the worldwide mining and minerals community through information
exchange and professional development. SME is the world’s largest association of
mining and minerals professionals.
Information contained in this work has been obtained by SME from sources believed to be
reliable. However, neither SME nor the authors guarantee the accuracy or completeness of
any information published herein, and neither SME nor the authors shall be responsible
for any errors, omissions, or damages arising out of use of this information. This work is
published with the understanding that SME and the authors are supplying information
but are not attempting to render engineering or other professional services. It is sold with
the understanding that the publisher is not engaged in rendering legal, accounting, or
other professional services. If such services are required, the assistance of an appropriate
professional should be sought. Any statement or views presented here are those of the
authors and are not necessarily those of SME. The mention of trade names for commercial
products does not imply the approval or endorsement of SME.
ISBN: 978-0-87335-363-2
Ebook: 978-0-87335-375-5
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Contents
Preface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
iii
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
iv CHALLENGES IN FINE COAL PROCESSING
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Contributors
H. Akbari Randy Corder
Southern Illinois University Arch Coal, Inc.
Department of Mining & Mineral Grafton, West Virginia
Resources Engineering
Carbondale, Illinois Van Davis
Alpha Natural Resources
Zulfiqar Ali Bristol, Virginia
Virginia Tech
Department of Mining & Minerals Greg DeHart
Engineering Arch Coal, Inc.
Blacksburg, Virginia Teays Valley, West Virginia
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
vi CHALLENGES IN FINE COAL PROCESSING
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
CONTRIBUTORS vii
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Preface
Coal mining and preparation have had a long history in the United States
and the world, serving as the engine of growth for many industries. Today,
new sources of energy, increased environmental awareness, and more stringent
regulations from the U.S. Environmental Protection Agency and other organi-
zations are changing the way coal is found, extracted, and used. As a result, fine
coal cleaning, dewatering, and refuse disposal are now at a major crossroads.
The increased level of fines, and near-density material in the inferior seams
being mined today, necessitates the development of more efficient fine coal
cleaning devices. This in turn requires improvements in traditional dewatering
techniques to address the need for acceptable moisture levels in plant prod-
ucts. Moreover, the larger volume of fine refuse being generated, coupled with
harsher disposal regulations, requires upgraded treatment options.
This book is a compilation of information presented at the 2012 Fine Coal
Symposium, sponsored by the Coal Preparation Society of America; the Pitts-
burgh Section of the Society for Mining, Metallurgy, and Exploration, Inc.;
and the Pittsburgh Coal Mining Institute of America.
Provided by international coal companies, major research organizations,
technology developers, and industry leaders, the information includes both
general knowledge and in-depth discussion on the current challenges facing the
industry, techniques for designing more efficient plants, and new cleaning and
dewatering technologies. The book is a practical yet cutting-edge resource for
plant designers, engineers, and other practitioners, and for university students
and faculty.
ix
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Perspectives on
Coal Fines 1
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Milestones in Fine Coal
Cleaning Development
Dave Osborne
ABSTRACT
Effective cleaning of fine coal (i.e., via the removal of non-coal material and subse-
quent reduction in moisture to an acceptable level) is mostly dependent on favor-
able economics (i.e., value of the product component obtained), capability, and
subsequent performance of the preparation equipment, and the extent to which
beneficiation of the total coal can be optimized.
The success of such an approach will be influenced by many other factors, not
the least being the proportion of fines in the raw coal. Hence, as mining has become
progressively more mechanized, the proportion of fines has increased and the jus-
tification for fine coal cleaning has tended to increase. However, the conundrum
associated with more fines is the added risk of increased moisture and the accompa-
nying need for improved and cost-effective dewatering equipment.
In fine coal cleaning, gravity methods such as jigs, tables, spirals, upward cur-
rent separators based on the forerunning Stokes classifier, as well as autogenous
cyclones are featured. Other non-gravity methods include froth and column
flotation and fines aggregation approaches including (oil) agglomeration and bri-
quetting. Perhaps the most desirable approach, dry cleaning, with its tremendous
appeal of avoiding the moisture issue altogether, is almost unachievable via conven-
tional methods, but some sorting, magnetic, and electrical techniques have offered
some promise over the years but are still far from being commercialized.
Not surprisingly, much of the development that has occurred in fine coal cleaning
has taken place with higher-value coal, often metallurgical, and almost all techniques
for both cleaning and dewatering have undergone some form of improvement or
development during the past decade as a result of better design or improved materials
of construction and wear resistance. Other focus areas have been process control and
monitoring, automation, and of course sampling and analysis.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
4 Perspective on Coal Fines
This chapter reviews the development pathway for fine coal beneficiation and
covers all aspects in a more general way, leaving the details for others to describe in
subsequent chapters.
INTRODUCTION
To provide some context for this chapter, several pertinent questions are first
addressed:
1. What is fine coal cleaning? The term coal fines is almost always taken to
mean raw or cleaned coal below 1 mm from which ultrafine material
sized below ~50 µm has been removed by some form of desliming
(classification) step. Cleaning of this “fraction” can be achieved by
application of any one of a variety of separators, predominantly in
water, to recover a valuable component(s). In some cases, this outcome
is achieved via a single stage and others in two or more stages.
2. What are its main drivers? Economics is the main driver and this can
be influenced by the prevalent financial climate. In many cases in past
eras, there has been no such justification for recovering “value” from
this fraction, and this coal was wasted only to be recovered at a later
stage when the financial climate became favorable. Other drivers can
include sustainable development, environmental impact, and new
technical developments.
3. When did fine coal cleaning start? Probably the earliest commercial
developments date back to the flowing-film concentration devices
introduced in the early 19th century, but serious coal cleaning whereby
a fine coal cleaning circuit was designed into a coal preparation plant
occurred in the 1930s. By this time, most of the gravity concentra-
tion devices (i.e., jigs, tables, cyclones, etc.) were in use, many having
emerged from mineral processing applications for which they were
originally developed.
4. What has its development pathway been? Fine coal beneficiation whereby
both cleaning and dewatering were incorporated probably emerged
as an integral part of flowsheet design during the 1940s and 1950s.
Knowledge transfer from the minerals industry applications undoubt-
edly contributed much to the development of both cleaning and dewa-
tering equipment, but there were some processes specifically introduced
for coal cleaning such as the Dutch State Mines (DSM) dense-medium
cyclone and water-only cyclone designs, and more recently the reflux
classifier albeit with some mineral processing influence.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Milestones in Fine Coal Cleaning Development 5
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
6 Perspective on Coal Fines
1800 1850
1500 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020
significantly with the result that coal preparation engineering work became
scarce and many experts migrated to other industries. However, with the intro-
duction of the Pentium chip, the engineering activity and storage of knowledge
and experience was transformed. By the mid-1990s, these mini-computer-
based systems became too expensive against the advancement of PC operating
systems, and with it the number of people needed to do engineering work of all
kinds was drastically reduced.
Finally, into the late 1990s and into this century, the transformation has
virtually become the norm and has translated into all facets (i.e., plant design,
construction, scheduling, supervisory control and data acquisition [SCADA]),
and perhaps the greatest benefit from all of this has been seen in the improved
treatment and cost of fine coal beneficiation.
CONTRIBUTING TECHNOLOGIES
Figure 1 is a time chart illustrating the pathway for the development of fine
coal beneficiation, which includes key milestones that have occurred during
this period.
The intent of this chapter is to set the scene for the comprehensive treatise
on fine coal beneficiation that follows in this book. The sections covered in
this chapter provide details of the various types of beneficiation and associated
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Milestones in Fine Coal Cleaning Development 7
equipment, including related aspects; and, for convenience, these are subdi-
vided into the following:
• Water-based separation
• Chemical separation
• Dry separation
• Magnetic and electrical separation
• Dewatering
• Computers and process control
• Materials science and engineering
Water-Based Separation
The earliest forms of separation employed flowing-film, sedimentation, or jig-
ging mechanisms to effect the separation.
Coal Jig
The use of jigging to create stratification and thereby clean coal is probably
among the oldest of coal cleaning techniques. Jigs in their simplest form have
been around from as early as the 15th century, when hand-operated versions
consisted of a framed sieve manually pulsed in a water tub. The first mecha-
nized coal jigs were recorded from as early as the 1840s, after which evolution
over the next 50 years focused on the jigging action, which changed from
moving a sieve frame to using a piston, then a diaphragm to pulse the medium
through a static sieve. In all cases, the raw coal was fed to the jig either after slur-
rying in water or after a step to “deslime” the feed, but in most cases there was
no further attempt to recover or beneficiate the fines.
By 1891, the father of the modern-day jig, the Baum jig, was developed
in Germany by Fritz Baum. This jig used air to displace water from one sealed
chamber, forcing the water through the jig bed in an adjacent chamber. The
use of air allowed for a more versatile control on the jig cycle over the older
diaphragm- or piston-driven Jeffrey and Harz jigs. With the introduction of
the Baum jig to Great Britain’s coal industry in 1903 and the Americas in 1928,
both the capacity and utilization of the jig rapidly grew until the 1950s when it
was felt that the capacity of the Baum jig had been reached with a 2.5-m-wide
unit. Wider units later appeared following the inclusion of the rotary air valve
that allowed the bed width of the Baum jig to be increased to 5 m (Lyman
1994).
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
8 Perspective on Coal Fines
In 1958, Takakuwa of Japan invented the Tacub jig (later replicated as the
Batac) by placing the air header directly below the bed, allowing bed widths of
up to 6 m. It was also in the 1950s that the innovation of using a mineral feld-
spar as “ragging” allowed the processing of finer sized coal by means of both of
these jig types. The modern-day versions of these jigs are capable of processing
material down to ~0.1 mm; however, there is a rapid deterioration in separation
efficiency below 0.5 mm, and this approach is not often employed when the
cleaned fines have little or no commercial value (Osborne 1988a). The modern
Batac jig is capable of processing up to 600 t/h of –10 mm + 0.5 mm coal.
More recent developments more specifically applied to fines treatment have
been the so-called in-line pressure jig and another variant, the Kelsey jig, the
latter of which incorporates centrifugal motion (Falconer 2003). Both employ
a ragging material and, although developed for mineral applications, so-called
“slurry jigs” have been seriously considered for fine coal cleaning applications
with various degrees of success reported (Bhattacharya 2009).
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Milestones in Fine Coal Cleaning Development 9
Settling Cone
Feed Water to Hoyois Troughs
Slurry Screens
Raw Slack
Elevator
Re-wash
Shale
Raw Slack Shale
Sump Sump
Overflow
Overflow
Main Pump
Sump
Figure 2 Hoyois trough washer, circa 1935 (Colliery Eng. Ltd., later Head Wrightson)
Shaking Tables
The modern wet shaking tables were originally derived from the continuous
belt concentrators used in the mid-1800s. In these early separators, the belt
concentrator used a flowing film to wash the light fraction (coal) down a belt,
leaving behind the minerals and shales. The first step toward current shaking
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
10 Perspective on Coal Fines
tables was taken by Hartwig in 1860 with the development of a vibrating end-
less belt (based on the mineral vanner). Further development by Linkenback,
with the circular stationary table—Anaconda-Evans with multiple round table
and Campbell with the bumping table—brought the dawn of water-based shak-
ing tables to the coal preparation industry, with the first bumping table instal-
lation in 1890. It was between 1896 and 1898 that Arthur Wilfley developed
the first differential motion table, which is still used for cleaning fine coal. The
Massco table was the name given to the original Wilfley table that was specifi-
cally designed for coal. Since then, many brands of shaking tables, all originally
designed for concentrating ores, have been marketed for coal cleaning. In 1918,
the most renowned coal cleaning table was introduced by Deister-Overstrom.
Since then, many hundreds of installations have been built, and probably the
most significant changes to the Deister table have been the incorporation of
multiple decks to a single unit, significantly increasing the capacity per floor
area, and also the improved material used for the table surface. By 1976, there
was an estimated 2,690 Deister tables operating in the United States. The
modern coal Deister tables are capable of processing feed between 0.15 mm
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Milestones in Fine Coal Cleaning Development 11
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
12 Perspective on Coal Fines
Courtesy of FLSmidth.
Water-Only Cyclones
The earliest cyclones were used to reclaim and thicken solids out of a slurry
material. Just prior to World War II, operators at the Maurits mine in The
Netherlands observed that when cyclones became plugged, the overflow
predominantly contained clean coal. From this observation, beneficiation
cyclones were developed by technologists at DSM. Since then, there has been
a significant amount of research into cyclones, resulting in increased capacity,
wear life, and optimization of performance. The basic design of the water-only
cyclone has not radically changed since the inception of the technology and is
typified by the wider cone angle (≥60°) compared with the classifying cyclone.
This smaller slope appears to encourage stratification at its surface, thereby con-
centrating the denser material. Several variants have emerged since the DSM
version, one of which is shown in Figure 4.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Milestones in Fine Coal Cleaning Development 13
Water-only cyclones have typically been used to process relatively fine mate-
rial, usually sized between 0.1 mm and 4 mm. The most common applications
have been twofold: reduction of pyrite from high-sulfur raw coal and effective
cleaning of fine coal with below 10% of near-gravity material. Because of the
relatively low single-unit efficiency, water-only cyclones have often been success-
fully used in two stages with a variety of configurations depending on coal type.
Dense-Medium Cyclones
Dense- (heavy-) medium cyclones were developed from the same beginnings as
the water-only cyclone, though more development was required to control the
dense medium before it emerged as a useful technology. This probably explains
why the first DSM dense-medium pilot plant was not built until 1945 (Sokaski
et al. 1968). Figure 5 shows the original flowsheet developed by researchers
at DSM. The first commercial-scale dense-medium cyclone plant specifically
intended for fine coal cleaning was built in Belgium in 1957 by DSM-Stami-
carbon and Evence Coppee (Mengelers and Dogge 1979). It was another 9
years before the second dense-medium plant was built, also in Belgium, but it
was not until 1961 that the normal dense-medium cyclone design was first used
in the United States, perhaps due to the availability of low-cost, readily easily
cleaned coal that was available in the United States at that time. Since then,
the use of dense-medium cyclones for coarser coal has grown considerably, but
the application of fine dense-medium cyclones has not enjoyed the same level
of adoption despite some confident attempts in the United States as well as in
Europe. Once again, there has been significant work into the cyclone with the
most progress for the dense-medium cyclones involving optimizing the separa-
tion, medium grades, spigot design, and retention of the dense medium.
All forms of fine coal cleaning have advanced in performance as a result of
the introduction of improved screening (or sieving), but perhaps none more
so than dense-medium cleaning where the introduction of the sieve bend,
shown in Figure 6, was perhaps a major breakthrough in enabling accurate siz-
ing and improved medium recovery in DSM circuits. Its application was more
widely utilized in other fine coal circuits, and innovations such as rapping or
induced vibration were introduced in attempts to overcome distribution and
blinding problems. Some circuits introduced tandem arrangements and others
used sieve bends to pre-scalp fines before multi-slope (banana screens) ahead
of dense-medium cyclone circuits (Leeder et al. 1986). Other innovative fine
coal screens have included the Stack Sizer by Derrick Company in the United
States, which introduced a high-frequency fines screen with a unique vibrator
in the 1960s from which has evolved several types for various applications. A
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
14 Perspective on Coal Fines
Raw Coal
Effluent
Figure 5 Flowsheet for the DSM cyclone using magnetite medium, circa 1945
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Milestones in Fine Coal Cleaning Development 15
Feed Inlet
Screen Deck
Underflow
Overflow
Despite the fact that much attention has been paid to the potential of the
dense-medium cyclone for fine coal cleaning, after almost 50 years of deter-
mined effort with at least a dozen commercially operated plants being built, the
process has yet to become the commercial proposition many thought would be
inevitable. The various attempts have been described elsewhere (Lathioor and
Osborne 1984) and are summarized in Table 1.
Since this publication, new plants have been built in South Africa with
radically improved medium recoveries, one of the major reasons why the earlier
plants failed, which has raised hopes that the application will meet with com-
mercial success and restore confidence. The numbers in Table 2 provide a clear
indication as to why this process is so appealing. Several others researchers have
fresh ideas, and a new generation of dense-medium cyclone circuits for cleaning
fine coal could well emerge within the next decade.
Chemical Separation
Flotation
Flotation of fine coal, like most other fine and ultrafine coal cleaning processes,
had its origins in mineral processing, and therefore, early flotation equipment
was adapted for coal applications with almost immediate success, primarily
because coal is a naturally hydrophobic material (Bury et al. 1921). Early tech-
niques were froth flotation, now referred to as conventional flotation, involving
cells agitated by a variety of mechanisms and differing tank designs offered by
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Table 1 Published performance data for fine coal dense-medium cyclone plants
16
2. Cyclone Type H&P H&P Krebs H&P DSM DSM Krebs DSM DSM DSM DSM Velmet DSM
3. Cyclone Diameter 102 203 610 203 610 610 254 350 350 250 250 250 150 250 150
(mm)
4. Feed Inlet Pressure 70 70 122 62 140 140 152 140 min 140 min 150 150 150 150 156 150
(kPa)
5. Probable Error (Ep) 0.055 0.028 0.065 0.053 0.149 0.188 0.024– 0.059 0.1 — 0.06 0.075 0.01 0.025 0.045 0.035/?
0.081
6. Imperfection (I) 0.108 0.08 0.071 0.156 0.169 0.028 0.041– 0.094 0.12 — 0.115 0.095 0.1 0.052 0.064 0.07/?
0.103
7. Cut-Point Density 1.51 1.35 1.92 1.34 1.88 1.73 1.59– 1.63 1.83 — 1.52 1.79 2 1.48 1.69 1.5/1.8
(d50) 1.79
9. Clean Coal Ash (%) — 1.6 7.9 5 6.6 10.3 — 5.3 6.8 8.2 13.3 8.3 12 13.7 7.4 9.9 7.5/15.0
11. Grain Size Range 0.30/0.15 0.34/0.044 0.6/0.15 1.0/0.15 0.6/0.15 0.6/0.15 0.6/0.1 0.3/0.75 0.3/0.15 1.0/0.3 0.3/0.15 1.0/0.1 1.0/0.1 1.0/0.1 0.5/0.075 0.5/0.075 0.5/0.15
(mm) (considered)
12. Feed Size Range 0.30/0.15 0.34/0.044 9.5/0 1.0/0.15 9.5/0 9.5/0 0.6/0.1 0.75/0 0.75/0 — — — — — 0.5/0.075 0.5/0.075 0.5/0.15
(mm) (to cyclones)
Perspective on Coal Fines
13. Magnetite Losses — 2.7 — — — — 1.5 — — — — 2.5 2.5 2.5 1.3 2.5 Max 1.5
Total kg/t feed
14. Type of Facility Test Plant Modular Test Plant Plant Plant Plant Plant Test Pilot Plant Test Pilot Test Plant Plant
Plant Plant Plant
15. Test Data 1974 1978 1981 1978 1979 1980 1958 1966 1966 1978 1978 1980
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
16. Comment ZnCl2 Magnetite Clean to 0 — Clean to 0 — Operated Until Rewash
Medium Medium Concept Concept 1983 Low Den-
sity/High
Density
Milestones in Fine Coal Cleaning Development 17
W-O
Cyclone Cyclone Cyclone Cyclone Cyclone Cyclone Spirals Reflux TBS
Size, mm 3 × 0.5 2 × 0.5 3×1 1×6 3.35 × 0.2 1 × 0.1 1 × 0.1 3 × 0.5 3 × 0.5
Epm 0.043 0.079 0.021 0.031 0.026 0.054 0.15 0.101 0.104
SG50 1.56 1.75 1.31 1.38 1.35 1.74 1.80 1.63 1.67
Epm/SG50 0.028 0.045 0.016 0.022 0.019 0.031 0.083 0.062 0.06
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
18 Perspective on Coal Fines
Column flotation for coal applications arrived in the late 1980s (Miller
1988). Since their introduction, applications have spread to all metallurgical
coal-producing countries due to the inherent ability to recover ultrafine coal
down into the submicrometer sizes. Each manufacturer’s design differs, with
the bubble generator perhaps being the most significant identifying feature.
One of the earliest columns was developed in Canada, acting on the need for
processing mineral ores and utilizing the bubble column concept that origi-
nated early in the 20th century in chemical industries. The development path
saw companies like Deister introduce the Floataire column, and universities
develop various cells, including the Ken-Flote column from University of
Kentucky Center for Applied Energy Research, the Microcell from Virginia
Polytechnic Institute and State University, and the Heyl and Patterson Miller
Cyclo-Cell in the United States; the Jameson cell from the University of New-
castle in Australia; and the Wemco/Leeds column from the University of Leeds
in the United Kingdom.
More recent developments have included introducing different forms of
microbubble design or improved variants of those currently used (i.e., porous
venturi tube, in-line mixer, turbo-generator, and various forms of spargers).
Most need fresh water to function effectively, but the Microcell uses slurry
recirculated from the bottom of the column, thereby creating a form of retreat-
ment of coal microparticles that might otherwise have been lost to the tail-
ings. Features like this will be important as applications in the preparation of
micronized coal for ultraclean coal applications such as coal-water slurry fuels
are commercialized. Combined with fine grinding mills such as high-pressure
rolls or stirred-ball mills, microflotation cells will probably be the separator of
choice to produce ultra-low ash coal (<0.5%) in such circuits. Directly injected
diesel engines (DICEs) and gas turbines are the likely users of these fuel types,
and interest appears to be growing, especially in China.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Milestones in Fine Coal Cleaning Development 19
SPS and Olifloc processes in Europe, Central Fuel Research Institute’s process
in India, and IPTACCS (integrated pipeline, transportation and coal cleaning
system) by BHP in Australia (Mehrotra et al. 1983) all had varying degrees of
success. However, all were impacted by the fluctuating oil price and improved
flotation techniques to the extent that there are still no commercial plants in
operation. The Otisca T-Process was developed for producing ultra-clean coals
with the ultimate objective of producing coal-water fuels. Most of the other
processes also anticipated this application, but some (BHP’s pipeline process
and NRC’s process) also intended the process to beneficiate valuable metal-
lurgical coal types for coke-making applications.
Briquetting, however, has been commercialized for many years and is
regarded as a mature technology. In 1848, a patent was granted to William
Easby for a briquetting process, but it took more than 50 years to emerge as
a commercial process. The coal briquetting process, as it ultimately evolved in
the United States, consisted of first drying the coal, then crushing and screen-
ing it, mixing the dry coal with about 6% molten asphalt binder, briquetting
this mixture in roll-type briquette machines, and finally cooling the briquettes
on a conveyor before loading them into railcars or diverting them to stockpile.
More than 6 Mt of coal briquettes were produced annually in the United States
before the process was hit by cheap oil and gas just after World War II.
The briquettes made by this process were used primarily for domestic heat-
ing, and many attempts were made to eliminate the asphalt binder, as the smoke
from the binder was the major user objection to the product. Coal briquetting
today has more definite applications. Coal is briquetted as an initial step in the
production of activated carbon.
There is undoubtedly growing interest in briquetting coal recovered from
tailings ponds or stockpiles of abandoned fine coal piles. Excessive amounts of
fine coal that have resulted from mining or crushing for greater liberation cannot
readily be shipped without some form of aggregation to a larger size to ensure that
a moisture specification is achieved. Briquetting is also used in the production of
form coke and has advantages in the production of metallurgical-grade coke as
well as in the coke-making process itself. Current machines, similar to the one
shown in Figure 7, can operate on coal feeds up to ~30 metric tons/h. So coal,
which launched a major briquetting industry in the first half of last century, may
well come full circle in the first half of this century (Komarek 1991).
Dry Separation
Mechanical dry separators for cleaning coal have been around since the early
1900s, with the first pneumatic oscillating machine for cleaning bituminous
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
20 Perspective on Coal Fines
coal introduced in 1916 and the first air dense-medium separator in 1923. Just
as with water-based cleaning, the pneumatic (dry) separators can be split into
three technologies: jigs, tables, and dense-medium separators.
All of the dry technologies appear to have emerged around the same time
(first half of the 20th century) and have undergone numerous developments at
different times since then, although no great breakthrough in their overall use,
let alone in fine coal treatment, has yet materialized.
The air jig is reportedly capable of separating coal down to 0.5 mm, whereas
the dense-medium vessels and pneumatic tables are only capable of separating
down to 0.8 mm and 3 mm, respectively. In the period between the 1930s and
the 1960s, as mining became increasingly more mechanized, dry separators
were more widely accepted, and several were incorporated into the design
of coal preparation plants, many as an initial step to “de-stone” run-of-mine
(ROM) coal.
The culmination of several factors challenged more widespread use of
mechanical dry separators between the late 1960s and early 1990s. These
included
• Inherent inefficiencies of the various technologies,
• Mechanized mining resulting in greater mining dilution in the feed
and a finer mean size of the ROM coal,
• General deterioration in the quality of available resources,
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Milestones in Fine Coal Cleaning Development 21
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
22 Perspective on Coal Fines
especially in arid areas where coal resources are abundant. Hence, the opportu-
nity to incorporate optimal coal cleaning across the supply chain exists, and the
coal supplier could actually contemplate supplying pulverized fuel to the power
plant at an attractive cost.
Additionally, tribo-electric beneficiation has other possibilities that include
the production of ultra-low ash coal (<1% ash) for injection metallurgy via the
generation of coal-based carbon for use as high-quality active carbon, or for
micronized refined coal for coal-water slurries for use in DICEs or coal-based
fuel cells.
Dewatering
In the beginning, fine coal was not recovered; and where the total raw coal
was washed, the raw fines were invariably rejected as tailings via settling cones
and thickeners. Hence, coal operations worldwide created coal-rich tailing
ponds and lagoons, the recovery of which would ultimately become of interest
when their value started to increase in the last two decades of the 20th century.
Another driver was the desire to close coal preparation plant water circuits, and
this would be most effectively achieved by introducing dewatering equipment
into the plants.
Mechanized mining and perhaps a greater awareness of the potential value
of coal fines, via enhanced petrography and research and development (R&D)
in utilization (particularly in the metallurgical coal market), were probably the
other major drivers in seeking to recover this fraction. New metallurgical coal
deposits with unprecedented fines contents, such as those in the Rocky Moun-
tain areas of Western Canada in the 1970s, probably added impetus to the need
for improved dewatering (Butcher et al. 1979). Because of the greater surface
area involved, the moisture retention properties of this fraction added signifi-
cantly to the moisture content and difficulty in handling the combined prod-
uct. This in turn drove the development of classifying cyclones, vacuum and
pressure filters, fine coal centrifuges and fine screens, together with improved
thickening and settling and filtration chemicals and even thermal drying.
So-called screen-scroll basket centrifuges, a relatively recent development,
are now widely used for dewatering small coal (–2 + 0.1 mm) from spirals or
teetered-bed separators. They have a conical filtering basket for drainage, while
an internal scroll with a different rotation speed assists in material transport
and determines the residence time of the coal. Higher centrifugal acceleration
is required for the finer material, so these centrifuges have higher speeds (maxi-
mum 700 rpm) and treat up to 100 metric tons/h of –2 + 0.1 mm coal produc-
ing about 12%–18% moisture. The onward trend is toward a larger basket for
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Milestones in Fine Coal Cleaning Development 23
the horizontal axis machines as per the European design, which reduces the
footprint and eases basket change and maintenance. Vertical axis machines,
which were preferred in North America, have the advantage of the coal being
presented to the scroll top, which provides full acceleration of the feed slurry.
The horizontal axis design feeds against the basket wall and is reported to pro-
duce marginally lower moisture contents at equivalent g-force.
Vacuum filtration, screen-bowl (including decanter-type) centrifuges, and,
to a lesser extent, hyperbaric and pressure filters are used to dewater fine coal
concentrate material. Perhaps the earliest entrants were the rotary vacuum disc
and drum filters. Early rotary disc filters, such as the Eimco Agidisc, became
the popular choice for flotation concentrates; and rotary drum filters, perhaps
more forgiving of coal type, were often used for more-difficult-to-filter fine
coal. Vacuum drum filters were popular in the 1990s, and some are still used
with string, belt, or snap-blow cake discharge options. A new breed of large-
diameter, high-capacity disc filters developed for the alumina industry in the
1970s were adapted and introduced into the coal industry for flotation con-
centrate dewatering.
The horizontal belt types have emerged as a popular choice in recent years
as they provide ease in operation, a high level of reliability, and are the most
flexible vacuum filters to adsorb feed material changes. Their top-fed filtration
principal uses the gravity effect and, therefore, if fed correctly, a layer of coarser
coal settles onto the filter cloth, forming a drainage bed, easing cake release, and
improving filtrate clarity.
The higher driving force of pressure filters compared to vacuum types
provides faster filtration kinetics and lower final product moisture, but it also
incurs higher installation and operating costs. Thus, pressure filtration is nor-
mally applied for difficult-to-treat ultrafine coal tailings or for the recovery of
“lost fines” from screen-bowl centrifuges. Pressure filtration is also sometimes
applied to ultrafine coal slurries for briquetting and agglomeration.
Variants to those previously mentioned include the hyperbaric filter, which
is an enclosed disc filter, installed inside a pressure vessel so that the dewater-
ing driving force is compressed air pressure. They provide up to twofold solids
throughput capacity compared to vacuum filters and up to 40% lower product
moisture for coal concentrates. A heating component can also be applied to
further enhance dewatering potential.
Another interesting variant is the Centribaric filter, which combines cen-
trifugation and pressure filtration within one process to substantially reduce
moistures over what can be achieved using conventional dewatering systems.
Ultrafine coal can be dewatered to as low as 20% by mass by using this type
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
24 Perspective on Coal Fines
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Milestones in Fine Coal Cleaning Development 25
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
26 Perspective on Coal Fines
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Milestones in Fine Coal Cleaning Development 27
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
28 Perspective on Coal Fines
mining company Rio Tinto to facilitate the transition of this new technology
to mine operations, which is the current Rio Tinto iron-ore focus.
In the processing area, treatment plants already featuring many control
features are an obvious target area, where the other major emphasis is safety.
The simple fact that those are areas of the plant which are almost always messy
and/or less accessible demonstrates that reaction times needed to avoid this
outcome are inadequate, and, hence, in the fine coal areas of the process circuit,
many opportunities exist for introducing automation and robotics solutions.
FUTURE DIRECTION
When the milestones are reviewed, clearly there have been very few new ben-
eficiation technologies emerging in the 21st century—just better designs of
existing technology. This is enhanced in many cases by the potential introduc-
tion of new materials, leading to improved maintenance and longer component
life, and the implementation of better control and design, in turn leading to
improved operation and greater efficiency. Further improvements in both areas
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Milestones in Fine Coal Cleaning Development 29
High
Life-Cycle
Partnerships
Integrated
Solutions R&D Co-development
Cooperation
Service/Maintenance Consulting Operating
Spare Parts Support Models
Growth in
Importance Mechanical
Automation
Components
Systems
Electrical
Components
Value Cone
Low
remain a major R&D challenge (Osborne 2010), and in this regard there is still
“lots to do.”
Some examples to whet the appetite are listed here:
• Efficient, economical coal dewatering of “microfine” coal (–0.015 mm)
• Effective cleaning of poorly floating ultrafine coal (0.150 × 0.045 mm)
• Accurate size classification at 0.045 mm or lower
• On-line coal characterization and species recognition in fine coal
streams
• Commercialization of high-capacity, high-resolution dry sorting
technologies
• “Whole of supply-chain” beneficiation solutions (e.g., pit to pulver-
ized fuel)
• Power and water conservation activated by cost of carbon and legal
liability
• Robotic sampling and on-line analysis of slurries and pastes
• Handling and possible utilization of ultrafine reject material (clays,
sands, sulfides, etc.)
The chapters that follow will no doubt address some of these challenges.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
30 Perspective on Coal Fines
ACKNOWLEDGMENTS
The author thanks colleagues Kelly Walton and Mark Jackson, coal prepara-
tion engineers with Xstrata Coal in Queensland, for their assistance in the
preparation of this chapter and other colleagues for proofreading and valuable
comments for improving the final version.
Opinions expressed in the chapter are the author’s own views and are not
necessarily shared by his colleagues or others at Xstrata Coal.
REFERENCES
Arrowsmith, G.H. 1957. Possible changes in coal preparation. In 2nd Symposium on
Coal Preparation. Paper 1. University of Leeds. October. pp. 11–30.
Bethell, P.J. 2007. Coal preparation: Current status and the way ahead. Presentation
to the Report Committee of the National Commission on Energy Policy—Coal
Study, Denver, CO, September 17.
Bethell, P.J. 2011. Personal communication.
Bhattacharya, S. 2009. Thermal coal preparation in India: Drivers, barriers, prospects
and challenges. South African Coal Processing Society. IV Biennial Coal Prepara-
tion Conference. Secunda, RSA.
Bury, E., Broadbridge, W., and Hutchinson, A. 1921. Froth flotation as applied to
washing of industrial coal. Trans Inst. Min. Eng. 110:243.
Butcher, S.G., Osborne, D.G., and Walters, A.D. 1979. Applications of computer
technology and special analytical methods developed for the design, commission-
ing and operation of plants treating highly friable Canadian coals. In IX Interna-
tional Coal Preparation Congress, Proceedings. Donetsk, Ukraine. Paper G3. pp.
145–160.
CSIRO. 2009. Mine automation. Earthmatters 19(March/April). Available: www
.csiro.au/files/files/pp19.pdf.
Dennis, P.I. 1998. Optimum processing of 1 mm by zero coal. Ph.D. dissertation. Vir-
ginia Polytechnic Institute and State University, Blacksburg, VA. pp. 1–3.
Driessen, G. 1945. The use of centrifugal force for cleaning fine coal in heavy liq-
uids and suspensions, with special reference to the cyclone washer. J. Inst. Fuel.
XIX(105):33–45.
Easby, W. 1848. Method for Converting Fine Coal into Solid Lumps. U.S. patent,
Aug. 29.
Falconer, A. 2003. Gravity separation: Old technique/new methods. Phys. Sep. Sci.
Eng. 12(1):31–48.
Firth, B.A. 2008. The Intelligent Plant—Measurement Requirements in Fine Coal
Cleaning and Dewatering Circuits. ACARP Project. Report C11069. p. 93.
Firth, B., and O’Brien, M. 2010. Fine coal measurement needs for improved control. In
XVI International Coal Preparation Congress, Proceedings. Edited by R. Honaker.
Littleton, CO: SME. pp. 679–688.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Milestones in Fine Coal Cleaning Development 31
Galvin, K.P., Callen, A., Spear, S., Walton, K., and Zhou, J. 2010. Gravity separation of
coal in the reflux classifier: New mechanism for suppressing the effects of particle
size. In XVI International Coal Preparation Congress, Proceedings. Edited by R.
Honaker. Littleton, CO: SME. pp. 345–351.
Hillman, J. 2003. Upward current washers. In A History of British Coal Preparation.
Edited by J. Hillman. United Kingdom: The Mineral Engineering Society. pp.
141–161.
Hoyois, L. 1935. GB456770 patent. Chaussee de Ransart, Gilly, Belgium, May 16.
Keles, S., Luttrell, G., Yoon, R.H., Estes, T., Schultz, W., and Bethell, P. 2010. Devel-
opment of the CentribaricTM dewatering technology. In XVI International Coal
Preparation Congress, Proceedings. Edited by R. Honaker. Littleton, CO: SME.
pp. 488–495.
Know, G. 1918. The Draper washer. Proc. Soc. Wales Inst Eng. 34:291.
Komarek, K. 1991. Binderless briquetting of peat, lignite, sub-bituminous and bitumi-
nous coals in a rolls press. In 22nd Biennial Conference of the Institute for Briquet-
ting and Agglomeration, Proceedings. San Antonio, TX. 22:233–242.
Lathioor, R.A., and Osborne, D.G. 1984. Dense medium cyclone cleaning of fine coal.
In 2nd International Conference on Hydrocyclones. Cranfield, UK: BHRA.
Leeder, W.R., Hogg, J.W., Jacobs, E.M., and Osborne, D.G. 1986. Application of high
capacity multi-slope screens for coal de-sliming applications in heavy media plants.
In X International Coal Preparation Congress, Proceedings. Edmonton, Canada.
1:300.
Lein, L. 2011. Mining’s new future: How the industry will change in the next decade.
Min. Eng. 63(February):41–46.
Luttrell, G.H., Honaker, R.Q., Bethell, P.J., and Stanley, F.L. 2007. Design of high effi-
ciency spiral circuits for coal preparation plants. In Designing the Coal Preparation
Plant of the Future. Edited by B.J. Arnold, M.S. Klima, and P. Bethell. Littleton,
CO: SME. pp. 73–87.
Lyman, G.J., and Davis, J.J. 1994. Cleaning of Coarse and Small Coal. Coal Beneficia-
tion—Part 7. Monograph Series. Australian Coal Preparation Society. 3:47–59.
Mehrotra, V.P., Sastry, K.V.S., and Morey, B.W. 1983. Review of oil agglomeration tech-
niques for processing of fine coals. Int. J. Miner. Process. 11(3):175–201.
Mengelers, J., and Dogge, C. 1979. A new technique for the treatment of 0.1mm fine
coal by means of heavy media cyclones. In VIII International Coal Preparation
Congress, Proceedings. Donetsk, USSR. Paper B1. pp. 104–111.
Miller, K.J. 1988. Novel flotation technology—A survey of equipment and processes.
In Industrial Practice of Fine Coal Processing. Edited by R.R. Klimpel and P.T.
Luckie. Littleton, CO: SME. pp. 347–363.
Osborne, D.G. 1986. Fine coal cleaning by gravity methods: A review of current prac-
tice. Coal Prep. 2:207–242.
Osborne, D.G. 1988a. Coal Preparation Technology. Vol. 2. London: Graham and
Trotman. pp. 276–285.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
32 Perspective on Coal Fines
Osborne, D.G. 1988b. Coal Preparation Technology. Vol. 2. London: Graham and Trot-
man. pp. 961–999.
Osborne, D.G. 2004. Coking coal and iron ore—Comparisons and contrasts. Arthur
LePage Lecture to the Australian Coal Preparation Society Conference, Pokolbin,
October.
Osborne, D.G. 2010. Value of R&D in coal preparation development. In XVI Inter-
national Coal Preparation Congress, Proceedings. Edited by R. Honaker. Littleton,
CO: SME. pp. 845–857.
Palowitch, E.R., Deurbrouck, A.W., and Parsons, T.H. 1991a. Hydraulic Concentra-
tion. In Coal Preparation. 5th ed. Edited by J.W. Leonard. Littleton, CO: SME.
pp. 414–416.
Palowitch, E.R., Deurbrouck, A.W., and Parsons, T.H. 1991b. Hydraulic Concentra-
tion. In Coal Preparation. 5th ed. Edited by J.W. Leonard. Littleton, CO: SME.
pp. 437–438.
Sokaski, M., Geer, M.R., and Yancey, H.F. 1968. Chapter 10: Dense medium separa-
tion. In Coal Preparation. 3rd ed. Edited by J.W. Leonard and D.R. Mitchell. New
York: AIME. p. 10-3.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Dealing with the Challenges
Facing Global Fine Coal
Processing
Peter J. Bethell
INTRODUCTION
This chapter covers some of the most important challenges facing fine coal
preparation internationally as we move forward in a dynamic, ever-changing
global energy scene. The challenges covered include
• Processing ever more poorly liberated coals for quality-constrained
markets,
• Plant water shortages,
• High reject coals, and
• Pyrite-rich coals.
Processing, dewatering, and waste disposal strategies to deal with these chal-
lenges will be tabled.
THE CHALLENGES
Poorly Liberated Coals
Most well-liberated coals in the coal-producing regions of the world have
already been mined. This leaves major challenges in providing high-quality
products from coal having more near-gravity material and very little low-
density, low-ash “cream.” Coals in India, Mozambique, and South Africa for the
most part will require substantial crushing. This will provide material liberated
enough to generate a marketable product for coking coal or high-heat steam
coals at economical yields. Crushing down to a top size of 12 mm, and maybe
even finer, will probably become commonplace. The impact of changing top
size from 50 mm to 12 mm can be seen from the washability curves generated
for an Indian coking coal (see Figure 1). Massive yield increases occur, which
33
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
34 Perspective on Coal Fines
50
45 50 mm Topsize
12 mm Topsize
40
35
30
% Ash DB
25
20
15
10
0
0 10 20 30 40 50 60 70 80 90 100
% Yield
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Challenges Facing Global Fine CoaL Processing 35
Feed Inlet
Lamella Chamber
Overflow
Access Door
Underflow Valve
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
36 Perspective on Coal Fines
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Challenges Facing Global Fine CoaL Processing 37
The use of on-line nuclear analyzers will become more important as the
quality of mined coals deteriorates and exacting quality specifications must be
met from inferior coals. Major changes in product ash levels for minor density
changes will be a standard problem. Plant density control circuits will need to
be maintained at a very high level in dealing with coals possessing high per-
centages of near-gravity material. Low tolerances for density variation will be
essential to maintain coal product quality. Briquetting or agglomeration may
be necessary in view of the large percentage of fines in the finished product.
Water-Constrained Plants
Most of the coal-producing areas of the world have water supply issues. Coal
processing plant design will need to be innovative and resourceful to conserve
water supplies to ensure project viability.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
38 Perspective on Coal Fines
Alternative strategies are present to deal with lack of water. These include
closed water circuits for conventional plant circuit design, “coarse wash-only
plants,” and dry cleaning circuitry. Desalination plants are also being intro-
duced to provide plant water sources, significantly pushing up treatment costs.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Challenges Facing Global Fine CoaL Processing 39
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
40 Perspective on Coal Fines
material, which traditionally served thermal coal markets unwashed, has dete-
riorated to the point that its quality raw is now unacceptable to the customer.
Examples of this phenomenon include the Western United States where coal
seams mined in Colorado and Utah have become progressively thinner and
higher in ash, delivering run-of-mine (ROM) products no longer salable on a
raw basis.
In these cases and in similar circumstances in some South African mines,
washing of the coarse coal will only provide sufficient beneficiation to render
an unsalable ROM coal salable. Typically in these plants, ROM coal is screened
at between 6 and 9 mm. The oversize material passes to a dense-medium bath
with the fines bypassed to product.
Conventional screening techniques will not provide efficient separation
dry at fine sizes. Consequently, innovative screen technology such as roller
screens and “flip flop” screens have been utilized to achieve high levels of
screening efficiency.
Recent plants incorporating this circuit design concept include Mid-
delburg in South Africa (Bivitec screen), which uses a Larcodems to clean
the coarse material. The Castle Valley (Roxon roller screen) and the Taggart
Global–built West Elk (Bivitec screen) plants of Arch Coal in the United
States use Peters vessels for coarse coal cleaning.
In both of the latter plants, dry screening is achieved on the novel screen
technology with the oversize (±6–8 mm) treated in a dense-medium bath cir-
cuit at high-efficiency levels. The Castle Valley circuit is shown in Figure 9 (Kel-
ley 2008). Figure 10 shows the West Elk plant. The flow in the plant from the
Bivitec screens to the pre-wet screens and after that the dense-medium bath fol-
lowed by appropriate screening, crushing, and drying can be seen in Figure10.
Wet fines generation and fine refuse disposal needs are kept to a minimum,
and, therefore, water requirements are also minimized, which is essential in the
dry, environmentally constrained Western United States.
The economic benefits of the coarse wash-only plants versus conventional
washing on the appropriate coals are enormous, both in terms of reduced
plant capital cost and increased plant yield. For the Western U.S. coals treated
in coarse wash-only plants, an advantageous feature of the coals has proved
to be the much higher ash level in the coarse material than in the fines. This
coupled with typically extremely well-liberated coarse material has enabled
these plants to produce a very salable combined product of washed coarse coal
and bypassed fines.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Challenges Facing Global Fine CoaL Processing 41
Roller
FEED
Screen
Magnetic Separator
Thickener Dewater
Screen
Centrifuge
CLEAN Pump
Conveyor COAL Conveyor
Pump COARSE
FINE Pump REFUSE
REFUSE
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
42 Perspective on Coal Fines
Dry Cleaning
The ability to have an efficient dry cleaning process for coals, particularly in
arid regions or in those lacking suitable water supplies or wet waste disposal
sites, is obvious. Air jigs and air tables, many resurrected from old designs,
have been developed to fulfill this need. Although the levels of processing
efficiency reached by these dry cleaning devices are considerably less than for
dense-medium processes, the ability to clean coal with no water requirements
has major benefits. To date, cleaning in the 50 mm × 6 mm size range has
proved successful, particularly where the devices have been used as a de-stoner
(de-shaler) reducing the ash level of previously unsalable coal to produce a
marketable product.
The FGX device (Figure 11) developed in China has found widespread
worldwide use in de-shaling (Orhan 2010). Work continues to drive the
efficient cleaning size for these units finer to increase their usage and enable
dirtier coals to be handled (Honaker 2010). For good dry cleaning to occur, it
is important that the feed material be reasonably dry. The FGX units operate
similarly to the wet shaking tables previously so common in plants of old. The
shaking action is supplemented by fluidizing air.
High-Reject Coals
Not only are we continuing to see coals with more near-density particles, we are
also encountering higher rejects in the coals feeding our plants. The reduction
in plant clean coal yield has an adverse impact on revenues. De-shaling ahead of
the conventional plant has been successfully applied in Germany and Australia
to maximize salable coal production by the use of Romjigs (Ziaja 2007).
Pyrite-Rich Coals
Dealing with pyrite-rich coals poses yet another challenge, particularly where
these coals have good coking characteristics. Density separating devices can
control sulfur levels within the constraints of the coal’s washability and the
density cut range of the device. Flotation, on the other hand, poses significant
additional challenges. Coal-associated pyrite on most occasions is very hydro-
phobic. Consequently, even if ash levels can be well controlled by the latest flo-
tation techniques, sulfur tends to be concentrated in the product giving a froth
concentrate much higher in sulfur than is acceptable in coking coal products.
This poses a new set of challenges, which will probably be addressed by complex
sizing/density separation/flotation circuitry.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Challenges Facing Global Fine CoaL Processing 43
T H E W AY A H E A D
Coal preparation in the future will face major challenges. The need to produce
a high-quality product will remain, but it will have to be done from low-yield,
middlings-rich coals. Water supplies will be harder to find and stricter environ-
mental regulations will be present. The positive side will undoubtedly be excel-
lent global demand for high-quality coking and steam products, particularly
from emerging markets such as India, China, and so forth.
Improving fine wet cleaning techniques such as spiraling and hindered-
bed separation as well as successfully driving down the bottom size treated by
DMCs will be needed. The ability to reduce spiral circuit cut points to the 1.5
specific gravity range would be extremely advantageous. Continual improve-
ment in techniques to effectively dewater the fines coal products from the new
plants will be required. Depending on handleability and dustiness, briquetting
and pelletizing processes may well be required to produce a similar product.
Improvements in dry screening efficiency and processing will be extremely
helpful in arid areas.
The area of ultrafine wet coal cleaning currently reserved for flotation (col-
umn and conventional) will need to be expanded to include efficient density
separation to process oxidized coals or those not efficiently treated by flotation.
Applications of devices such as the Falcon concentrator and Kelsey jig, modi-
fied to achieve efficient separation in coal cleaning, may well prove viable.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
44 Perspective on Coal Fines
Through innovation and a “can do” attitude, the coal processing engineers
of the future will rise to the challenges facing them as coal preparation engi-
neers always have. Adaptability will be the key to survival and success.
REFERENCES
Bethell, P., and Barbee, C. 2007. Today’s coal preparation plant a global perspective.
In Designing the Coal Preparation Plant of the Future. Edited by B.J. Arnold, M.S.
Klima, and P.J. Bethell. Littleton, CO: SME. pp. 9–20.
Bethell, P., and Luttrell, G. 2005. Effects of ultrafine desliming on coal flotation cir-
cuits. In Proceedings, Century of Flotation Symposium. Brisbane. p. 43.
Brodzik, P. 2007. Application of Derrick Corporation’s Stack Sizer in clean coal spiral
product circuits. In Designing the Coal Preparation Plant of the Future. Edited by
B.J. Arnold, M.S. Klima, and P.J. Bethell. Littleton, CO: SME. pp. 89–96.
Cresswell, G. 2010. Process design of the Phola Coal Preparation Plant. In XVI Inter-
national Coal Preparation Congress, Proceedings. Edited by R. Honaker. Littleton,
CO: SME. pp. 66–75.
Dekorte, D.J. 2002. Dense media beneficiation of fire coal revisited. In Proceedings of
the International Coal Preparation Congress. Johannesburg. p. 43.
Galvin, K., Walton, K., and Zhou, J. 2010. Gravity separation and classification of fine
coal using the hydrodynamics of inclined channels. In Thirteenth Australian Coal
Preparation Conference. p. 224.
Gupta, B.K., and Bethell, P. 2010. Deep cone thickener at the Lone Mountain plant.
In Proceedings of the XVI International Coal Preparation Congress. Edited by R.
Honaker. Littleton, CO: SME. pp. 674–678.
Hahn, J. 2011. Performance, operation and maintenance experience of coal ultrafines,
filtration with modern high speed disc filters. In Proceedings of the South African
Coal Processing Society Meeting, Sekunda.
Honaker, R., Luttrell, G., and Mohanty, M. 2010. Coal preparation research in the
USA. In Proceedings of the XVI International Coal Preparation Congress. Edited by
R. Honaker. Littleton, CO: SME. pp. 864–874.
Keles, S., Luttrell, G., Yoon, R.H., Estes, T., Shultz, W., and Bethell, P. 2010. Devel-
opment of the CentribaricTM dewatering technology. In Proceedings of the XVI
International Coal Preparation Congress. Edited by R. Honaker. Littleton, CO:
SME. pp. 488–495.
Kelley, M., and Bethell, P. 2008. The design commissioning and operation of the Castle
Valley plant. Presented at 2008 Annual SME Meeting, Salt Lake City, UT.
Kennedy, D., et al. 2006. Improvements in flotation column recovery using cell-to-cell
circuitry. In Proceedings of the 23rd Annual International Coal, Aggregate and Min-
eral Processing Exhibition and Conference, Lexington, KY. pp. 89–98.
Luttrell, G., Stanley, F., Honaker, R., and Bethell, P. 2003. Operating guidelines for coal
spirals. In Proceedings of the 20th Annual International Coal Preparation Exhibi-
tion and Conference, Lexington, KY. p. 69.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Challenges Facing Global Fine CoaL Processing 45
Orhan, E.C., Ergun, L., and Altiparmak, B. 2010. Application of the FGX separator
in the enrichment of Catalagzi coal: A simulation study. In Proceedings of the XVI
International Coal Preparation Congress. Edited by R. Honaker. Littleton, CO:
SME. pp. 562–570.
Ziaja, D., and Yannoulis, G.F. 2007. Is there anything new in coarse or intermediate
coal cleaning? In Designing the Coal Preparation Plant of the Future. Edited by B.J.
Arnold, M.S. Klima, and P.J. Bethell. Littleton, CO: SME. pp. 43–59.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Properties and Challenges
for Fines Processing
Barbara J. Arnold
ABSTRACT
Coal is used as a fuel as well as a source of carbon for metallurgical processes. It can
also be a source of carbon-based chemicals. For each of these markets, specifications
must be met for various coal properties. Generally, these are related to either the
organic (macerals and elemental analysis of carbon, hydrogen, etc.) or inorganic
(e.g., minerals or moisture) components of a coal, and coal rank, the increase in
coalification of a deposit of organic material overlays changes in composition—
higher carbon content and lower moisture, for example.
The organic and inorganic components also create opportunities and challenges
for the selection of fine coal processing technologies. This chapter reviews these prop-
erties and gives information related to fine coal processing options.
INTRODUCTION
According to the International Organization for Standardization (ISO 2005),
organic deposits of peat become “coal” when the moisture content is below
75%. In addition, the highest ash content recognized as coal is 50%. Of course,
some refuse piles containing greater than 50% ash are being reprocessed to
recover valuable coal products. The ash value and moisture are certainly impor-
tant coal properties associated with producing a salable product. But other
properties are also important—both physical and those related to the petrology
and mineralogy of the coal.
Physical properties like the size distribution of a coal are included in many
contracts. Some have a top size limit—perhaps 50 mm (2 in.) for coal used in
pulverized coal combustion systems or 6 mm (0.25 in.) for fluid bed systems.
Many contracts will limit the amount of fines that can be in a coal to limit
coal handling issues (Arnold 1995b). Moisture constraints, especially those
for surface moisture, can be considered a physical property. Limits on surface
47
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
48 Perspective on Coal Fines
moisture content are related to coal handling issues; and total moisture (surface
plus inherent) will impact the evaporative load on power plant boilers.
The petrologic and mineralogic properties are related to the organic and
inorganic components in a coal. Properties like carbon or hydrogen content—
rank-related phenomena—determine whether a coal can be used for metallur-
gical purposes. The composition of the minerals also influences the slagging or
fouling behavior of the resultant ash in a boiler.
These properties are also important when considering selection of coal pro-
cessing equipment. Of course, size distribution and washability will be impor-
tant for the selection of all coal cleaning equipment, including those for the
coarse coal fractions. However, as coal particles become finer, surface properties
become more important and associated minerals, such as clay minerals, play an
increasingly important role in equipment selection and chemical additive selec-
tion for processing and dewatering. Slurry properties will become increasingly
important as well. pH and the concentration of other elements like Ca2+ or
Mg2+ will influence the behavior of coal and mineral particles.
The froth flotation process has been described by many investigators, and Han-
sen and Klimpel (1985) noted that three components—chemistry, operational,
and equipment parameters—influence results. Mineralogy, which for coal includes
the petrologic macerals as well, was included under operational parameters. Other
operational parameters included feed rate, particle size, pulp density, and tempera-
ture. These same parameters will be important for any other fines processing device,
whether for beneficiation, classification, dewatering, or thickening devices.
And these operational parameters will influence the “chemistry” required
for efficient processing. For flotation, this can be collectors, frothers, activa-
tors, depressants, and pH control. For dewatering or thickening, chemistry will
include dewatering aids or flocculants, and pH control can also come into play.
And perhaps, attention to chemistry can play a role in classification with the
use of dispersants or pH control for dispersion. And one can ask the question
as to whether “chemistry” is required for processing coal in gravity separation
devices like water-only cyclones, spirals, reflux classifiers, and so forth. But the
“chemistry” that is required for a particular coal is based on its organic and
inorganic composition and the resulting slurry properties.
This chapter will delve into these coal properties—these operational
parameters—and the challenges they bring to fine coal processing.
COAL
No two coals are alike. Even within a given coal seam, coal properties vary. A
coal started out as a deposit of plant matter that underwent decomposition and
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Properties and Challenges for Fines Processing 49
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Table 1 Coal macerals—classification, origin, and properties
50
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
tia, hyphae remains
Source: Davis et al. 1976.
Coal Properties and Challenges for Fines Processing 51
Many other properties of coal also change with coal rank, as given in
Table 6. A number of these will impact processing options.
For example, the Hardgrove grindability index (HGI) increases up to about
the low volatile bituminous rank, with anthracite becoming more difficult to
grind. Knowing the HGI of a coal will give an indication of the amount of fines
that can be generated through handling and cleaning. In some cases, for even
coarse coal circuitry, very high HGI coals have been handled very carefully,
including gravity-fed dense-medium cyclone circuits to eliminate pumping the
coal to the cyclone. Fines circuit capacity should also be increased to accom-
modate these increases in fines. One source has indicated that projected fines
should be increased by at least 40% for very high HGI coals (Stanley 2007).
The oxygen content of coal decreases with increasing coal rank, as does the
amount of oxygen present as OH and COOH groups in and on the surface
of coal. The higher oxygen content is related to low hydrophobicity and poor
floatability (Sun 1954; Gutierrez and Aplan 1984). It is thought that the pres-
ence of reactive oxygen groups that form upon oxidation—like those of humic
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Table 4 ASTM classification of coals by rank
52
Fixed Carbon Limits, % Volatile Matter Limits, % Calorific Value Limits, Btu/lb / MJ/kg
(dry mineral matter-free basis) (dry mineral matter-free basis) (moist* mineral matter-free basis)
Agglomerating
Class/Group = or > < > = or < = or > < Character
Anthracite
Meta-anthracite 98 — — 2 — —
Anthracite 92 98 2 8 — — Nonagglomerating
Semiantracite 86 92 8 14 — —
Bituminous
Low volatile 78 86 14 22 — —
Commonly agglomerating
Medium volatile 69 78 22 31 — —
(there may be nonag-
High volatile A — 69 31 — 14,000†/32.6 — glomerating varieties with
†
notable exception in the
High volatile B — — — — 13,000 /30.2 14,000/32.6
high volatile C group)
High volatile C — — — — 11,500/26.7 13,000/30.2
10,500/24.4 11,500/26.7 Agglomerating
Subbituminous
Subbituminous A — — — — 10,500/24.4 11,500/26.7
Perspective on Coal Fines
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Lignite B — — — — — 6,300/14.7
Source: ASTM 2005.
Note: This classification does not include a few coals, principally nonbanded varieties, that have unusual physical and chemical properties and which come within
the limits of fixed carbon or calorific value of the high volatile bituminous and sub-bituminous ranks. All these coals either contain less than 48% dry mineral
matter-free fixed carbon or have more than 15,500 moist mineral matter-free Btu per pound (Btu/lb × 2,326 = kJ/kg).
* “Moist” refers to coal containing its natural inherent moisture but not including visible water on the surface of the coal.
† Coals having 69% or more fixed carbon on the dry mineral matter-free basis shall be classified according to the fixed carbon, regardless of calorific value.
Coal Properties and Challenges for Fines Processing 53
acids—cause this recovery loss. Highly oxidized coal and humic acid extracts
from an oxidized bituminous coal exhibited a large negative zeta potential
with essentially no point-of-zero charge (PZC) (Wen 1977). Firth and Nicol
(1981) studied coal flotation behavior in the presence of clay and humic acids.
They hypothesized that clay absorbed more oil in the presence of humic acids so
that there was less collector available to float the coal. Table 7 gives results from
Arnold and Aplan (1986b) showing that humic acids depress coal flotation in
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
54 Perspective on Coal Fines
70
50
30
Contact Angle, θ°
Captive Bubble
60
40
60
20
50
0.5 1.0
0
0 1.0 2.0 3.0 4.0 5.0
Mean Maximum % VRO
both the presence and absence of clay. pH also has an effect in this humic acid/
clay/coal system.
The PZC—the pH at which the surface charge is zero—also varies with
coal rank. Below the PZC, the coal surface is positively charged, and above the
PZC, the coal surface is negatively charged (see example in Figure 2). Most
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Table 6 Approximate values of some coal properties in different rank ranges (dry, mineral matter-free basis unless otherwise noted)
High Volatile Bituminous Bituminous
Sub-
Property Lignite bituminous C B A Medium Volatile Low Volatile Anthracite
Calorific value, moist mineral 7000 / 16.3 10 000 / 23.3 12 000 / 27.9 13 500 / 31.4 14 500 / 33.7 15 000 / 34.9 15 800 / 36.7 15 200 / 35.3
matter free, Btu/lb / MJ/kg
%C, range 66–76 71–81 76–84 79–85 83–89 86–92 87–92 92–97
Density (He), vitrinite ←Increases ←Increases ←Increases ←Increases Minimum Greatly increases→
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Surface area (CO2), m2/g
Hardgrove grindability index 35–50 40–55 45–80 45–80 45–80 70–115 70–115 25–50
Source: Aplan 1993 modified and expanded from the original tabulation of Given 1984.
55
56 Perspective on Coal Fines
Table 7 Effect of pH on Pittsburgh coal flotation in the presence of humic acid and
clay
Coal Recovery (0.1 kg/t MIBC), %
pH 3.5 pH 6.5
Clean coal alone
Tap water 84.7 87.7
Tap water +0.08 kg/t humic acid 68.9 56.4
Distilled water 82.5 80.6
Distilled water +0.08 kg/t humic acid 50.9 58.2
20% Kaolinite
Tap water 62.4 87.9
Tap water +0.08 kg/t humic acid 46.4 73.5
Distilled water 30.3 40.2
Distilled water +0.08 kg/t humic acid — 56.8
Source: Arnold and Aplan 1986b.
PZC results in the literature are reported as the surface charges in distilled
water—the absence of cations and anions. The presence of these ions, as in tap
water (or preparation plant water), will reduce the magnitude—either negative
or positive—of the surface charge.
As alluded to in Table 7 and in the discussion related to clay later in this
chapter, it will be shown that the presence of the ions affects the behavior of
the clay and coal in a flotation system. It will almost certainly have an effect on
fine coal dewatering and thickening and the reagents applied in these systems.
This coal/mineral/water system needs to be dewatered or thickened. In a
review of dewatering algorithms in the 1990s for the development of a flow-
sheet simulator sponsored by the U.S. Department of Energy and the Electric
Power Research Institute (Arnold 1995a, 1996, 1997, 1999), it was found
that important properties for predicting final moisture content included the
following:
• The average particle size using a mean diameter based on the volume.
That is, it uses both size and specific gravity distributions in the calcula-
tion. The amount of fines and ultrafines will be a big determinant of
the final moisture content.
• Bed or cake thickness.
• Dewatering force (pressure, gravity, vacuum, etc.) and dewatering time.
But also of importance were properties related to the slurry—surface
tension and liquid density for the screening algorithms; and surface tension,
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Properties and Challenges for Fines Processing 57
+40
+30
+20
+10
Zeta Potential, ζ, mv
–10
–20
–30
–40
–50
2 3 4 5 6 7 8 9 10 11 12
pH
viscosity of filtrate, and temperature for centrifuge and filter algorithms. A coal
property—contact angle (hydrophobicity)—was included in the screening
algorithm. The oxidation of coal or pyrite, ions in solution, and so forth, will all
play a role in the moisture content from dewatering devices as they effect these
properties. These were developed from work by Wakeman and Rushton (1977)
and Wakeman (1979).
Inorganic Components
Inorganic components in coal vary greatly. As-mined coal includes the mineral
impurities within the coal seam, those in partings, and those from the roof and
floor. These all become part of the feed to a coal preparation plant, and certain
of these minerals are found more prominently in the tailings from preparation
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
58 Perspective on Coal Fines
Calcite § 12 0–22 17
Pyrite § 4 1–11 2
Feldspar § § 30 <1
Dolomite § § 14
plants (O’Gorman and Walker 1972; Bradley et al. 1980a, 1980b). Table 8
summarizes the minerals present in U.S. coals and preparation plant tailings.
Inorganic ions are then present from the dissolution of these minerals or the
oxidation products of these minerals.
Minerals
The minerals in coal can be present as large, discrete, easily removed particles
or as finely disseminated particles that would require significant particle size
reduction to liberate the mineral from the coal. But mineral particles are
also present throughout the range of sizes of coal and in varying degrees of
liberation or locking. As given in Table 9, these coal–mineral particles have
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Properties and Challenges for Fines Processing 59
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
60 Perspective on Coal Fines
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Properties and Challenges for Fines Processing 61
+40
+30
+20
Zeta Potential, ζ, mv
+10
–10
–20
–30
-40
2 3 4 5 6 7 8 9 10 11 12
pH
+30
+20
Zeta Potential, ζ, mv
+10
–10
–20
–30
-40
2 3 4 5 6 7 8 9 10 11 12
pH
Figure 3 Zeta potential of kaolinite, illite, and bentonite in tap and distilled waters or
with cations present
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
62 Perspective on Coal Fines
SUMMARY
In summary, there are important coal properties to be considered when process-
ing coal, especially fine coal.
1. No two coals are alike.
2. Consider the macerals present in a coal. Vitrinite macerals are gener-
ally the predominant maceral found in a coal, but higher concentra-
tions of inertinite macerals like fusinite could result in lower flotation
yields than expected, as these macerals are less hydrophobic than the
accompanying vitrinite. This could also play a role in the dewatering
behavior of coal fines. And it could also explain differences noted in
the behavior of dewatering aids that seemingly work for some but not
all coals. Fusinite is also a more friable maceral and could generate
more fines than expected.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Properties and Challenges for Fines Processing 63
3. Consider coal rank. Higher rank coals are more hydrophobic. Lower
rank coals have higher oxygen contents and humic acids may be pres-
ent in solution in greater concentration. Higher rank coals are more
friable and more fines will be present.
4. Consider the minerals present in a coal. Pyrite is easily oxidized and
can contribute to low pH. Clay minerals should be characterized and
have differing behavior with pH, the presence of other ions in solution,
and the presence of humic acids.
Although liberated mineral particles may be easily rejected, their
dissolution will affect the ions in solution in preparation plant waters.
These ions will affect pH, the charge on coal particle surfaces, the
charge on the mineral particles (especially noting the two differently
charged surfaces for clay minerals), and certainly the utilization of
reagents for flotation, dewatering, and thickening. The final moisture
content in dewatering may also be affected by the water or slurry
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
64 Perspective on Coal Fines
REFERENCES
Aplan, F.F. 1993. Coal properties dictate coal flotation strategies. SME Trans. 294.
Arnold, B.J. 1985. Coal froth flotation: The effects of petrological and mineralogical
constituents. M.S. thesis, University Park, PA: Pennsylvania State University.
Arnold, B.J. 1995a. A Guide to Coal Handling. TR-105110. Palo Alto, CA: Electric
Power Research Institute.
Arnold, B.J. 1995b. Predicting the moisture content of coal dewatered by vacuum fil-
ters. In Proceedings of the Twelfth Annual International Pittsburgh Coal Conference.
pp. 455–460.
Arnold, B.J. 1996. Predicting the moisture content of coal dewatered by centrifuges.
SME Preprint 96-139. Littleton, CO: SME.
Arnold, B.J. 1997. Predicting the moisture content of coals dewatered by screens. SME
Preprint 97-128. Littleton, CO: SME.
Arnold, B.J. 1999. Simulation of dewatering devices for predicting the moisture con-
tent of coals. Coal Prep. 20:35–54.
Arnold, B.J., and Aplan, F.F. 1986a. The effect of clay slimes on coal flotation, Part I:
The nature of the clay. Int. J. Miner. Process. 17:225–242.
Arnold, B.J., and Aplan, F.F. 1986b. The effect of clay slimes on coal flotation, Part II:
The role of water quality. Int. J. Miner. Process. 17:243–260.
Arnold, B.J., and Aplan, F.F. 1989. The hydrophobicity of coal macerals. Fuel
68:651–658.
ASTM D388. 2005. Standard Classification of Coals by Rank. West Conshohocken,
PA: ASTM International. Available from www.astm.org/Standard/index.shtml.
Bradley, P.B., Hogg, R., and Aplan, F.F. 1980a. Mineralogical characterization of black-
water solids. Trans. AIME 268:1831–1836.
Bradley, P.B., Hogg, R., and Aplan, F.F. 1980b. Particle size consist of black-water sol-
ids. Trans. AIME 268:1836–1841.
Brown, D.J., and Smith, H.G. 1954. Continuous testing of frothers. Colliery Eng.
31:245–250.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Properties and Challenges for Fines Processing 65
Burdon, R.G., Booth, R.W., and Mishra, S.K. 1976. Factors influencing the selection
of processes for the beneficiation of fine coal. In Proceedings of the Seventh Inter-
national Coal Preparation Congress, Sydney, Australia.
Davis, A., Spackman, W., and Given, P.H. 1976. The influence of the properties of coals
on their conversion into clean fuels. Energy Sources 3(1):55–81.
Esposito, M.C., Chander, S., and Aplan, F.F. 1987. Characterization of pyrite from coal
sources. In Process Mineralogy VII. Edited by A.H. Vassiliou et al. Warrendale, PA:
TMS-AIME. pp. 475–493.
Firth, B.A., and Nicol, S.K. 1981. The influence of humic materials on the flotation of
coals. Int. J. Miner. Process. 8:239–248.
Given, P.H. 1984. An essay on the organic geochemistry of coal. In Coal Science, Vol.
3. Edited by M.L. Gorbaty, J.W. Larsen, and I. Wender. Orlando, FL: Academic
Press. pp. 63–252.
Gutierrez-Rodriguez, J.A., and Aplan, F.F. 1984. The effect of oxygen on the hydropho-
bicity and floatability of coal. Colloids Surf. 12:27–51.
Hansen, R.D., and Klimpel, R.R. 1985. The influence of frothers on particle size and
selectivity in coal/sulfide mineral flotation. SME Preprint 85-9. Littleton, CO:
SME.
ISO (International Organization for Standardization). ISO 11760:2005(E). 2005.
Classification of Coals. Geneva, Switzerland: ISO.
Jowett, A., El-Sinbawy, H., and Smith, H.G. 1956. Slime coatings of coal in flotation
pulps. Fuel 35:303–309.
Luttrell, G., McKeon, T., and Bethell, P. 2000. An in-plant evaluation of froth wash-
ing for conventional coal flotation circuits. SME Preprint 01-191. Littleton, CO:
SME.
Lynch, A.J., Johnson, N.W., Manlapig, E.V., and Thorne, C.G. 1981. Mineral and Coal
Flotation Circuits. New York: Elsevier. pp. 21–56.
Mishra, S.K. 1978. The slime problem in Australian coal flotation. In Australasian
I.M.M. Mill Operators Conference. Mt. Isa. pp. 159–168.
Neczaj-Hruzewicz, J., Szczypa, J., and Czarkowski, N. 1974. Influence of flota-
tion reagents on formation of gangue slime coatings on coal. Trans. I.M.M.
83:C261–263.
O’Gorman, J.V., and Walker, P.L. 1972. Mineral Matter and Trace Elements in U.S.
Coals. Report No. 61, Interim Report No. 2. U.S. Department of Interior, Office
of Coal Research.
Purcell, R.J. 1982. Unpublished data. University Park, PA: Pennsylvania State
University.
Stach, E., Mackowsky, M-TH., Teichmuller, M., Taylor, G.H., Chandra, D., and
Teichmuller, R. 1975. Stach’s Textbook of Coal Petrology. Berlin: Gebrüder
Borntraeger.
Stanley, F. 2007. Personal communication.
Sun, S.C. 1954. Effect of oxidation of coals on their flotation properties. Trans. AIME
199:396–401.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
66 Perspective on Coal Fines
Szczypa, J., Neczaj-Hruzewicz, J., and Sablik, J. 1973. Some properties of slime coating
in coal-gangue systems. Trans. I.M.M. 82:C167–169.
van Olphen, H. 1963. An Introduction to Clay Colloid Chemistry. New York: Inter-
science Publishers.
Wakeman, R.J. 1979. The prediction and calculation of cake dewatering characteristics.
Filtr. Sep. (Nov.-Dec.):655–669.
Wakeman, R.J., and Rushton, A. 1977. Dewatering properties of particulate beds.
J. Powder Bulk Solids Technol. 1(2):64–69.
Wen, W.W. 1977. Electrokinetic behavior and flotation of oxidized coal. Ph.D. thesis,
University Park, PA: Pennsylvania State University.
Yancey, H.F., and Taylor, J.A. 1935. Froth Flotation of Coal; Sulfur and Ash Reduction.
RI 3263. U.S. Bureau of Mines.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Maximizing Fine Pyrite
Rejection at the Arch Coal
Leer Plant
Cory Chafin, Peter Bethell, Greg DeHart,
and Randy Corder
ABSTRACT
Optimized fine pyrite rejection has been used by Arch Coal, Inc., at their Leer
plant to maximize coking coal recovery. Original circuit design produced 187 tph
(tons per hour) of minus 1 mm material that would be sold on the steam market.
By the redesign and optimization of the fines circuit it is anticipated that the Leer
plant will produce an increase of 149 tph of high volatile A coking coal and a net
reduction of 141 tph of middlings generating substantial revenue increases. Pyrite
concentration can be achieved by the use of classification and density separation
devices. A series of simulation and test programs has shown the benefit of pyrite
rejection through cyclone and sieve classifiers, as well as compound spiral concentra-
tors and reflux classifiers. The findings of these studies were used in the redesign of
the Leer Mining plant flowsheet. Additional sulfur reduction benefits are expected
when complete circuit optimization has occurred.
BACKGROUND
Year 2011 brought many changes to the coal business. One of these changes was
the Arch Coal, Inc. (ACI), acquisition of International Coal Group that was
completed in June 2011. Part of that acquisition was the Tygart #1 property,
which was in the process of being developed and is located in Taylor County,
W.V., near the town of Grafton. At the time of the acquisition, a flowsheet
had been developed and was in the very early stages of construction. The flow-
sheet was designed to clean coal from the Tygart #1 mine, which is mining in
the Lower Kittanning seam. The original flowsheet design envisaged placing
all minus 1 mm clean coal into a low realization steam market along with
67
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
68 Perspective on Coal Fines
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Maximizing Fine Pyrite Rejection 69
Table 1 1 mm × 150 μm raw coal washability data, which represents 14.42% of the
total sample
Direct Data Cumulative Recovery
Specific
Gravity Weight, % Ash, % Suflur, % Weight, % Ash, % Sulfur, %
1.35 Float 62.72 4.98 1.38 62.72 4.98 1.38
1.40 Float 5.21 13.64 1.71 67.93 5.64 1.41
1.45 Float 2.82 19.53 2.19 70.75 6.20 1.44
1.50 Float 1.86 27.22 2.73 72.61 6.74 1.47
1.60 Float 1.01 34.67 3.00 73.62 7.12 1.49
1.70 Float 1.37 42.25 3.50 74.99 7.76 1.53
1.80 Float 0.34 47.89 3.62 75.33 7.94 1.54
2.00 Float 0.78 54.25 3.82 76.11 8.42 1.56
2.20 Float 0.59 64.95 4.92 76.70 8.85 1.59
2.20 Sink 23.3 89.65 3.93 100.00 27.68 2.13
150 μm × 0 Circuit
Previous testing on nominal minus 0.6 mm (28 mesh) samples showed poor
sulfur rejection in flotation (Table 2) with concentrate sulfur levels in the
1.65–1.85% range when maintaining combustible recoveries greater than 80%.
Flotation feed was highly pyritic, and, unfortunately, the hydrophobic nature of
the pyrite made it highly problematic. This sulfur level was obviously unaccept-
able in a coking coal product where the rest of the plant circuits would have had
to be operated at low separation densities (i.e., <1.50 SG) to make an acceptable
overall sulfur specification (nominally <1.2%).
Release analysis testing on a pit bottom coal sample of minus 150 µm mate-
rial also showed very high product sulfur levels of >2.5% (see Table 3). A more
detailed review of the raw coal washability data showed that the high sulfur
levels in the flotation concentrate were almost certainly due to pyritic sulfur. In
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
70 Perspective on Coal Fines
Table 2 Results for flotation of minus 28 mesh Lower Kittanning seam coal
Product Reagent Drops Weight, % Ash, % Sulfur, %
7% Solids Loading
Tails 80.02 18.29 3.23
Concentrate 2 19.98 7.4 1.15
Tails 44.83 26.97 4.52
Concentrate 4 55.17 6.98 1.41
Tails 15.41 56.61 8.23
Concentrate 6 84.39 8.93 1.66
Tails 11.82 65.4 9.83
Concentrate 8 88.18 9.63 1.77
Tails 10.63 68.78 9.92
Concentrate 10 89.37 10.36 1.85
10% Solids Loading
Tails 75.55 18.38 3.23
Concentrate 2 24.45 9.98 1.39
Tails 46.08 25.91 5.34
Concentrate 4 53.92 8.4 1.47
Tails 23.33 39.86 6.47
Concentrate 6 76.67 8.94 1.63
Tails 17.78 47.57 7.73
Concentrate 8 82.22 10.06 1.82
Tails 13.61 62.22 9.58
Concentrate 10 86.39 9.02 1.63
fact, the organic sulfur of this coal is typically 0.75% (see Table 4). A review of
the raw coal washability showed excellent liberation of the pyrite into the finer
fractions of the feed stock.
Simulation modeling of the classifying cyclones in the plant showed that
liberated pyrite with 5.0 SG would be classified at a much finer size than the
coal and rock with 50% and 95% passing sizes of 35 µm and 45 µm, respectively,
for the pyrite. If this cyclone overflow stream was fed alone to the flotation cir-
cuit, feed and concentrate sulfur levels would be much less. In-plant testing has
also been conducted to confirm this simulation. Testing around a bank of clas-
sifying cyclones with a coal feed similar to that of the Leer washability showed
a major reduction in cyclone overflow sulfur levels (Table 5). Sulfur levels were
reduced from 2.28% in the feed to 1.11% in the overflow, with 5.60% sulfur in
the underflow of the minus 150 µm fractions.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Maximizing Fine Pyrite Rejection 71
Table 3 Flotation release analysis of minus 100 mesh Lower Kittanning seam coal,
representing 8.14% of the sample*
Direct Values† Cumulative Values†
Sulfur, Sulfur,
Increment Wt, % Ash, % % CR‡ Wt, % Ash, % % CR‡
Concentrate 1 37.29 8.59 2.66 67.70 37.29 8.59 2.66 67.70
Concentrate 2 7.38 10.79 2.55 13.08 44.67 8.95 2.64 80.78
Concentrate 3 2.01 15.26 3.16 3.38 46.68 9.23 2.66 84.16
Concentrate 4 3.49 24.01 3.31 5.27 50.17 10.25 2.71 89.43
Concentrate 5 1.48 31.82 3.00 2.00 51.65 10.87 2.72 91.43
Tailings 2 1.61 76.28 4.71 0.75 53.26 12.85 2.78 92.18
Tailings 1 46.74 91.58 0.59 7.82 100.00 49.65 1.76 100.00
* Collector: diesel; frother: MIBC.
† Dry basis.
‡ CR = combustible recovery.
The minus 150 µm material in the cyclone underflow in the plant would
pass through a cleaning device, a desliming device, and then would typically
report to froth flotation. A review of the ultrafines sulfur levels from the
aforementioned plant shows some sulfur removal in the spiral circuit for the
ultrafines (i.e., 5.65% in the feed down to 2.63% in the product). However,
this 2.63% sulfur material passes through the desliming sieve (see Table 6) and
would ultimately end up in the flotation circuit, resulting in elevated sulfur
levels.
Testing performed on minus 150 µm material on compound spirals has
shown great promise in reducing pyritic sulfur levels. Figures 1 and 2 represent
an Illinois plant in which spiral clean coal sieve effluent samples were treated on
compound spirals and then with flotation or treated in spirals or with flotation
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
72 Perspective on Coal Fines
Wt, Ash, Sulfur, Wt, Ash, Sulfur, Wt, Ash, Sulfur, Partition
Size Fraction % % % % % % % % % Coefficient
+1.5 mm 0.99 35.41 1.75 1.47 35.27 1.90 0.00 0.00 0.00 100.00
1.5 × 1 mm 5.73 21.69 2.03 9.27 25.38 2.63 0.00 0.00 0.00 100.00
1 × 0.25 mm 46.89 18.75 2.21 64.81 18.84 2.37 0.60 9.68 1.19 99.66
0.25 × 0.15 mm 10.05 18.33 2.73 12.41 20.26 3.33 2.79 3.95 0.85 92.27
+0.15 mm 63.66 19.21 2.27 87.96 20.00 2.52 3.39 4.96 0.91 98.58
0.15 mm × 11.52 18.79 2.76 8.43 31.94 4.54 17.71 5.01 0.96 56.08
75 μm
75 × 45 μm 5.38 21.30 3.29 2.03 52.24 7.64 13.56 8.13 1.08 28.65
45 μm × 0 19.44 28.74 1.71 1.58 54.09 9.09 65.34 30.56 1.15 6.09
Composite 100 21.13 2.27 100.00 22.20 2.90 100 22.13 1.10 72.84
0.15 mm × 0 36.35 24.49 2.28 12.04 38.27 5.66 96.62 22.73 1.11 25.05
*Based on % solids.
Wt, Ash, Sulfur, Wt, Ash, Sulfur, Wt, Ash, Sulfur, Partition
Size Fraction % % % % % % % % % Coefficient
+1 mm 9.47 9.49 1.31 13.43 9.69 1.52 0.00 0.00 0.00 100.00
1 × 0.25 mm 68.76 7.92 1.34 77.14 7.74 1.31 18.72 8.10 1.24 91.94
0.25 × 0.15 mm 8.39 7.71 1.27 6.53 7.77 1.35 39.17 10.71 1.59 31.58
+0.15 mm 86.62 8.07 1.33 97.10 8.01 1.34 57.89 9.87 1.48 82.28
0.15 mm × 0 13.38 24.06 2.63 2.90 15.29 2.05 42.11 23.74 2.79 16.01
Composite 100 10.21 1.50 100 8.22 1.36 100 15.71 2.03 73.46
*Based on ash.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Maximizing Fine Pyrite Rejection 73
100
90
80
70
60
Yield, %
50
40
30
20
Flotation Only
10 Spiral Only
Spiral + Flotation
0
0 10 20 30 40 50
Ash, %
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
74 Perspective on Coal Fines
100
90
80
70
60
Yield, %
50
40
30
20
Flotation Only
10 Spiral Only
Spiral + Flotation
0
0 2 4 6 8
Sulfur, %
100
90
Cumulative Yield, wt %
80
70
60 Pre-Spiral Release Sulfur
50
Post-Spiral Release Sulfur
40
Pre-Spiral Rate Sulfur
30
20 Post-Spiral Rate Sulfur
10
0
1 1.2 1.4 1.6 1.8 2 2.2 2.4
Cumulative Sulfur, wt %
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Maximizing Fine Pyrite Rejection 75
cyclones were chosen because of the superior pyrite rejection they provide.
The nominal 1 mm × 150 µm material along with a large portion of the minus
150 µm pyrite-rich material will be treated in primary/secondary reflux clas-
sifiers. The primary cleaning devices in this circuit will make the coking coal
separation operating at approximately 1.50 SG. The reject from the primary
reflux classifiers will be retreated by secondary devices where the middlings
material will be recovered. The primary and secondary reflux classifier products
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
76 Perspective on Coal Fines
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Maximizing Fine Pyrite Rejection 77
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
78 Perspective on Coal Fines
CONCLUSION
Because of the concentration and isolation of liberated pyritic sulfur in the
minus 1 mm material at Leer Mining, a large percentage of fine coal will be
moved from a low-realization steam market for which it was initially destined
to a higher-realization coking coal market. The anticipated results of the rede-
signed fines circuit will take 187 tph originally destined for the steam market
due to high ash and sulfur levels, and put 149 tph and 46 tph on the coking
coal and steam markets, respectively, generating substantial revenue increases.
This has been achieved by selecting equipment for the redesigned circuits that
will allow for maximum concentration and removal of ultrafine pyritic sulfur
from the coking coal circuitry. The concentrated pyrite is then treated through
compound spirals where the differential density of pyrite and coal is exploited
prior to secondary column flotation. This provides further potential enhance-
ments to the coking coal yield. The ability to run lower densities and therefore
lower sulfur levels has also been exploited by the incorporation of reflux classi-
fiers. Additional levels of sulfur rejection are expected when full plant process
is achieved and circuits can be fully optimized.
REFERENCES
Baumharth, T., Bethell, P.J., and Gupta, B. 2005. Recovering an additional 20 tph
of coal through a deslime column flotation circuit addition at Lone Mountain
Processing–Virginia. In Proceedings, Coal Prep AGG 2005. Stamford, CT: Penton
Media. pp. 39–50.
Bethell, P.J. 2010. Arch Coal processing philosophy, east and west. In Proceedings, XVI
International Coal Preparation Congress. Edited by R. Honaker. Littleton, CO:
SME. pp. 1–8.
Bethell, P.J., and DeHart, G. 2006. The design, construction, and commissioning of
the new 2000 tph Arch Coal Cardinal preparation plant. In Proceedings of the XV
International Coal Preparation Congress and Exhibition. Beijing: China National
Coal Association. pp. 79–88.
Bethell, P.J., and Luttrell, G. 2005. Proceedings, Centenary of Flotation Symposium. Bris-
bane, Australia, June 5–9. The Australasian Institute of Mining and Metallurgy.
Galvin, K.P., Callen, A.M., and Spear, S. 2008. Extending the size range of the reflux
classifier. ACARP Project C16040.
Honaker, R.Q., Boaten, F., and Luttrell, G.H. 2007. Ultrafine coal classification using
150mm gMax cyclone circuit. Miner. Eng. 20:1218–1226.
Yoon, R.H., and Luttrell, G.H. 1994. Microcel column flotation scale-up and plant
practice. CMP Proceedings of the 26th Annual Meeting. Paper 12.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Technology
Developments and
Plant Installations 2
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
StackCell Flotation—A New
Technology for Fine Coal
Recovery
Michael Kiser, Robert Bratton, Gerald Luttrell,
Jaisen Kohmuench, Eric Yan, Lance Christodoulou,
Van Davis, and Fred Stanley
A bstract
During the past decade, column flotation cells have become widely accepted for
the upgrading of fine coal streams. This popularity can be largely attributed to the
ability of columns to remove high-ash clays from the froth product via the addition
of wash water to a relatively deep froth. Although there are numerous successful
column installations, discussions with both end-users and engineering firms have
identified certain design criteria that can make these installations challenging.
The greatest of these challenges is the overall size of the cells and the associated
foundation loads. To address this problem, a new high-intensity flotation system
known as the StackCell has been developed. This technology makes use of pre-
aeration coupled with a high-shear feed canister. This arrangement provides effi-
cient bubble–particle contacting, thereby substantially shortening the residence
time required for coal collection and virtually eliminating most of the column
height. This article reviews the design features of this innovative technology and
presents recent data obtained from full-scale installations.
I ntroduction
Column flotation has become the dominant method of recovering fine frac-
tions in the coal industry. The use of columns has led to increased metallurgi-
cal performance when compared to that of mechanical flotation cells. This
improvement in product quality has been proven by comparing plant flotation
data to a release analysis curve (Dell et al. 1972). Studies have also been per-
formed that show how the use of column flotation affects the bottom line of
81
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
82 Technology Developments and Plant Installations
a plant (Luttrell et al. 1999; Kohmuench et al. 2004; Baumgarth et al. 2005).
These studies report that the plants benefit from an increase in overall yield
because of the improvement in product grade from the flotation circuit. This
increase in product grade is linked to the application of wash water used in
column flotation. This countercurrent flow of water is applied to the froth and
minimizes the non-selective recovery of high-ash ultrafine material that is nor-
mally hydraulically entrained in the froth of conventional flotation machines.
Column flotation does have its own set of design challenges though. The
first of these challenges is simply the size of the column. The cell must be tall
to achieve the desired residence time and minimize internal mixing, which
can be detrimental to cell performance. This design minimizes the plant floor
space required for the cell but increases the foundation loads. The large size of
the column also leads to difficulties with fabrication and installation of column
cells. The economics associated with plant design typically lean toward fewer
large-diameter cells. The largest diameter cell that can be shipped in the United
States as a single piece is 4.5 m (15 ft). Larger cells can be installed, but these
cells must be shipped in multiple sections and require more on-site assembly.
Additionally, larger-diameter cells must be taller to maintain the proper aspect
ratio, at least 2:1, which then adds to the overall foundation load.
The design challenges previously mentioned show that there is a need for
a new generation of flotation machine. A machine is needed that is capable of
delivering column-like performance, while also improving some of the design
and operational challenges associated with column flotation. Based on experi-
ence gained over the last decade with the design, engineering, and operation
of coal flotation circuits, Eriez has developed a new flotation cell that offers
high capacity, reduction in both size and horsepower, and superior metallurgi-
cal performance. Although column flotation will be a requirement for some
applications, this new approach offers an alternative that provides column-like
performance with reduced capital, installation, and operating costs.
technology D E S C R I P T I O N
Figure 1 illustrates the working features of the StackCell technology. During
operation, feed slurry is introduced to the cell through a side (or bottom) feed
port. At this point, low-pressure air is added to the feed slurry. The aerated
feed slurry then travels into the aeration chamber where significant shear is
imparted to the system. The shear forces imparted to the system are used to cre-
ate bubbles for bubble–particle collisions. In fact, all of these bubble–particle
collisions occur in the aeration chamber prior to discharge into the outer tank.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
StackCell Flotation—A New Technology 83
Aeration/
Contacting
Chamber Wash-Water
Manifold
Froth
Launder
Tails
Feed Inlet
Outlet
Air Manifold
When the slurry enters the outer tank, phase separation occurs between the
froth and pulp. A pulp level is maintained in the outer tank to provide a deep
froth that can be washed to minimize the entrainment of ultrafine high-ash clay
material. The froth overflows into a froth collection launder, while the tailings
are discharged using either a control valve or mechanical weir system. The sys-
tem is specifically designed to have both a small footprint and a gravity-driven
feed system. This allows multiple units to be “stacked” in series on subsequent
levels in the plant or placed ahead of existing column or conventional flotation
circuits.
Why Multistage?
The enhanced performance made possible by the stacked arrangement can be
mathematically quantified using the standard tanks-in-series flotation model
(Lynch et al. 1981). According to this model, the recovery (R) of a given species
from a single well-mixed flotation tank can be estimated by
kτ
R= (EQ 1)
1 + kτ
where k is the flotation rate constant and τ is the residence time. The rate con-
stant represents how quickly particles float and is normally reported in units of
min–1 (i.e., mass floated per unit mass in the cell per unit time). This parameter
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
84 Technology Developments and Plant Installations
RN = Ri + Ri (1 − Ri ) + Ri (1 − Ri ) + Ri (1 − Ri )
2 3
(EQ 2)
+ + Ri (1 − Ri ) = 1 − (1 − Ri )
N N
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
StackCell Flotation—A New Technology 85
100
N=∞
54
3
90 2
N=1
80
Recovery, %
70
60
N
N
R=1–
N + kτ
50
0 1 2 3 4 5 6
kτ
Figure 2 Effect of residence time and number of cells in series on the recovery of
floatable material (assumes k = 0.8 min–1)
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
86 Technology Developments and Plant Installations
Feed
Combined Tails
Product
k ∝ CbC p E (EQ 4)
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
StackCell Flotation—A New Technology 87
According to Equation 1, the very high rate constant (k) created by the
high-shear environment within the aeration chamber allows the StackCell to
operate at a correspondingly lower residence time (τ) without adversely impact-
ing the recovery. Field studies conducted with a pilot-scale unit showed that a
residence time of less than 10 seconds was often adequate for good contacting
when using StackCell technology. Consequently, the required cell volume for
a StackCell installation is significantly less, thereby reducing both equipment
and installation costs. Structural steel requirements are considerably less due to
the reduction in tank weight and live load. For a typical installation, the overall
space requirement for the stacked-cell design is half the volume of an equivalent
column circuit. Shipping and installation are also simplified, because the units
can be shipped fully assembled and lifted into place, complete and without field
welding. Moreover, the energy input per unit ton processed is typically lower
for the StackCell, since energy is only expended for the purpose of creating
bubbles and for bubble–particle contacting, and not for particle suspension like
conventional flotation cells. In addition, the aeration chamber operates under a
near-atmospheric pressure in a manner that removes the need for a compressor
to overcome the hydrostatic or dynamic head. As a result, a low-pressure and
maintenance-friendly blower can be used as opposed to a compressor.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
88 Technology Developments and Plant Installations
I ndustrial E VA L U AT I O N
To demonstrate the performance capabilities of StackCell technology, a full-scale
unit was installed and commissioned at an industrial coal preparation plant. The
plant processed run-of-mine coals from several seams supplied by both under-
ground and surface mines. The StackCell unit consisted of a single 3.7-m- (12-
ft-) diameter cell equipped with a 76-cm- (30-in.-) diameter aeration chamber.
The single StackCell unit was installed as a scalping system ahead of two existing
flotation columns. Historical data suggested that the two column cells were often
overloaded because of plant production demands. The tailings stream from the
StackCell was equally split and fed to the two existing columns.
Figure 5 shows the impact of the StackCell installation on the combustible
recovery and refuse ash for the entire flotation circuit. For the first 149 samples
taken prior to the installation, the two-column cells provided an average recov-
ery of 74.4% and a combined refuse ash of 72.5%. After the installation, the
combined recovery for the StackCell and two-column cells improved to 83.7%
and the refuse ash increased to 80.7%. The increased recovery is significant
considering that less than 10% more cell volume was added to the circuit via
the installation of StackCell technology. In fact, the aeration chamber provided
an additional residence time of only about 5–10 seconds to the total flotation
circuit. More recently, the average monthly plant recoveries have increased to
more than 90% (i.e., 90.88%), while the average monthly tailings ash values
have increased to nearly 86% (i.e., 85.9%).
Close inspection of the test data indicates a gradual improvement in overall
performance since the StackCell was installed. The continued improvement
can be largely attributed to the optimization of operating variables such as
reagent dosage, froth depth, aeration rate, and wash-water addition rate that
occurred over time as a result of fine tuning by the plant operators. For example,
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
StackCell Flotation—A New Technology 89
100
80
Recovery, %
60
Avg = 74.4% Avg = 83.7%
40
20
0
100
80
Reject Ash, %
60
20
0
0 50 100 150 200 250 300
Sample Number
Figures 6 and 7 show the impact of optimization on the clean coal quality and
recovery for the +0.30 mm (+48 mesh), 0.30 × 0.15 mm (48 × 100 mesh),
0.15 × 0.045 mm (100 × 325 mesh), and –325 mesh size fractions. The high
ash content in the –325 mesh fraction was substantially reduced from about
43.4% to less than 13.3% once the froth washing system was optimized. This is
due to the elimination of entrained ultrafine, non-floatable high-ash material
in the –325 mesh fraction. Before the opitmization of the wash-water addition,
roughly 13% of the concentrate was made up of this entrained material. After
the wash water was optimized, roughly 2.84% of the concentrate was made
up of this high-ash material. The plant data continue to show that the quality
of the froth product is sensitive to froth depth and wash-water addition rate.
Therefore, it is important that these values be properly monitored and con-
trolled to optimum settings.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
90 Technology Developments and Plant Installations
50
Before Optimization
45
After Optimization
40
35
30
Clean Coal Ash, %
25
20
15
10
0
+48 48 × 100 100 × 325 –325
Size Class, mesh
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
StackCell Flotation—A New Technology 91
14
Before Optimization
After Optimization
Weight of Non-Floatable Material Present in Concentrate, % of Conc.
12
10
0
+48 48 × 100 100 × 325 –325
Size Class, mesh
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
92 Technology Developments and Plant Installations
9 9
8 8
7 7
6 6
Dilution Washes
5 5
Percent
2 2
1 1
0 0
0 5 10 15 20 25 30 35
Froth Depth, in.
Figure 8 Concentrate quality and dilution wash data for various froth depths
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
StackCell Flotation—A New Technology 93
12 12
Total Conc. % Ash
10 Non-Float –325% Weight of Total 10
Dilution Washes
8 8
Dilution Washes
Percent
6 6
4 4
2 2
0 0
0 100 200 300 400 500
Wash-Water Rate, gpm
Figure 9 Concentrate quality and dilution wash data for various wash-water
addition rates
The concentrate samples from the three non-zero rates tested were made
up of similar amounts of –325 mesh hydrophilic material ranging from 4.7%
to 5.6% of the total concentrate weight at the 78 m3/h (345 gpm) and the
91 m3/h (400 gpm) tests, respectively. Interestingly, the ash content of the
concentrate increased slightly at the highest wash-water rate. One possibility
for the higher value was that pressure variations due to pump cycling/surg-
ing in the preparation plant created fluctuations in the wash-water flow rate.
Although this could be a contributing factor to the high ash value, it is unlikely
that it is solely responsible since the other data points were subjected to the
same testing conditions. A more likely explanation for the unexpected increase
in ash is short-circuiting of the wash water into the concentrate. This possibil-
ity is supported by the fact that the dilution washes also did not increase as the
water rate was increased to the highest rate. Thus, more of the wash water must
have reported to the concentrate, which reduced the dilution washes for that
test. Further testing is suggested to determine whether this phenomenon is site
specific or an inherent characteristic of this particular flotation machine.
SUMMARY
A new high-capacity flotation technology, called the StackCell, has been devel-
oped as an alternative to both conventional and column flotation machines. This
technology makes use of pre-aeration and a high-shear aeration chamber that
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
94 Technology Developments and Plant Installations
R eferences
Baumgarth, T., Bethell, P., and Gupta, B. 2005. Recovering an additional 20 tph coal
through a deslime column flotation circuit addition at Lone Mountain Process-
ing–Virginia. In Proceedings, 22nd Annual International Coal Preparation and
Aggregate Processing Exhibition and Conference, Lexington, KY, May 2–5. pp.
41–50.
Dell, C.C., Bunyard, M.J., Rickelton W.A., and Young, P.A. 1972. Release analysis: A
comparison of techniques. Trans. IMM (Sect C), 81:787.
Kohmuench, J.N., Davy, M.S., Ingram, W.S., Brake, I.R., and Luttrell, G.H. 2004.
Benefits of column flotation using the Eriez Microcel. Tenth Australian Coal
Preparation Conference, Proceedings, Polkolbin, NSW, Australia, October 17–21.
pp. 272–284.
Kohmuench, J.N., Mankosa, M.J., and Yan, E.S. 2008. An alternative for fine coal Flota-
tion. Coal Prep. Soc. Am. J. 7(1):29–38.
Luttrell, G.H., Kohmuench, J.N., Stanley, F.L., and Davis, V.L. 1999. Technical and
economic considerations in the design of column flotation circuits for the coal
industry. SME Preprint No. 99-166. Littleton, CO: SME.
Lynch, A.J., Johnson, N.W., Manlapig, E.V., and Thorne, C.G. 1981. Mineral and Coal
Flotation Circuits. New York: Elsevier Scientific. pp. 44–55.
Stanley, F., King, P., Horton, S., Kennedy, D., McGough, K., and Luttrell, G. 2006.
Improvements in flotation column recovery using cell-to-cell circuitry. In 23rd
International Coal Preparation Exhibition and Conference, Proceedings, Lexington,
KY, May 1–4. pp. 87–98.
Williams, J.J.E., and Crane, R.I. 1983. Particle collision rate in turbulent flow. Int. J.
Multiphase Flow 9(4):421–435.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing
Developments in Anglo
American Thermal Coal
South Africa
Chris Swanepoel
ABSTRACT
This chapter details the approaches adopted in fine coal processing within Anglo
American Thermal Coal South Africa, outlining the journey followed since 1995.
This historical review of fine coal processing focuses on beneficiation, dewatering,
and agglomeration with particular reference to (1) flotation, detailing the devel-
opment of Multicell technology and the subsequent evolution to Dual cell; (2)
successful dewatering of ultrafine thermal coal by achieving specified low moistures
and ensuring product handleability; and (3) briquetting of ultrafine material
resulting in the creation of a salable product, which adds value and minimizes
environmental liability.
95
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
96 Technology Developments and Plant Installations
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing Developments 97
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
98 Technology Developments and Plant Installations
BACKGROUND
Flotation
The process of froth flotation, which separates minerals according to the surface
properties, has been in existence since 1860, and flotation is used in the treat-
ment of fine ground ores. The extent to which minerals are separated is known as
recovery, and the selectivity of the process is defined by grade, which means that
the recovery is defined as the mass ratio of the valuable mineral in the concentrate
to the valuable mineral in the feed. Grade is defined as the ratio of the mass of the
valuable mineral in a processing stream to the total mass of that stream.
Water is used as a medium within the cell in which more dense particles
would sink. As the solid particles are suspended in solution (i.e., by means of
mechanical agitation), a second process, based on the surface properties (i.e.,
the particles are either hydrophobic or hydrophilic), takes place at the same
time. This process requires air to be either injected into the system or intro-
duced by agitation of the slurry, which produces bubbles within the slurry.
Hydrophobic particles attach to the bubbles and are floated to the surface
of the slurry as a result of the air being less dense than water. For a particle to
stay attached to a bubble, the gravitational force experienced by the particle
must not overcome the adhesion to the bubble.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing Developments 99
Coal Flotation
To distinguish between coal flotation and other mineral flotation processes, it
is necessary to consider that in other mineral processes, the material is specially
ground to be selectively processed with flotation, but in the case of coal, the
process is designed to generate the least amount of fines possible. This means
that the fines arise from various processing stages, which could have been
exposed to different environments, and, as such, the properties may also differ
throughout the processing stream.
When considering coal, the combustible mass is considered to be the valu-
able mineral and ash constitutes the gangue mineral. The flotation process is
utilized in industry, because flotation is the only process for effectively and
economically beneficiating –150 µm coal.
Flotation Reagents
Even though coal is naturally hydrophobic, the use of collecting reagents when
processing coal increases the floatability of the coal by altering the behavior of
coal within the process. Bubbles have a natural tendency to break the air–water
interface as it reaches the water surface, which is not a desired effect in the
flotation process, because the valuable minerals need to be removed from the
cell before the interface is destroyed. Therefore, a substance called a frother is
used to facilitate air dispersion into fine bubbles and increase froth stability.
This keeps the material floating above the pulp surface long enough for froth
removal.
Dewatering/Filtration
Slurry (i.e., a mixture of solids and water) is typically dewatered/deslimed using
sieve bends, static screens, vibrating screens, hydrocyclones, and centrifuges.
The slimes fraction (i.e., typically <150 µm) poses a great challenge to the
industry. Generally this fraction is disposed of by making use of slurry ponds,
which allows settling of the material and minimum reclamation of water back
to the operation. This, however, creates various environmental liabilities for the
industry and therefore a major global drive to maximize water recovery effec-
tively minimizing raw water consumption and slimes disposal. At this point in
time, filtration is the most common method of dewatering, and by which the
typical types of equipment used are horizontal belt filters, disc filters, drum
filters, and plate-and-frame filter presses.
Other than direct filtration, froth flotation can also be used as an upstream
process, in which case a salable high-value product can be produced. This
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
100 Technology Developments and Plant Installations
product is then filtered and blended with the typical export product. However,
given that the product contains moisture, the lowest possible product moisture
is required. Additionally, flotation tailings are produced, which can then be
filtered in the same manner to maximize overall water reclamation.
Filtration is the removal of solid particles from a fluid, by passing the fluid
through a filtering medium on which the solids build up. Many factors can be
important in selecting a filtration process, but because mineral processing oper-
ations are concerned primarily with recovering the solids at large throughputs,
the selection of equipment is considerably narrowed. Filtration equipment size
is specified by the surface area necessary to produce the required product. As
with sedimentation, particle properties cannot be adequately measured, and
small-scale filtration tests must be carried out to obtain basic data. Filters can be
operated in two basic modes. Constant pressure filtration maintains a constant
pressure so that the flow rate falls slowly from a maximum at the start of the
cycle. Most continuous filters can be considered to operate on this principle,
using a vacuum to provide the pressure difference. Constant rate filtration
requires gradually increasing pressure as the cake builds up and increases the
resistance to flow. A common approach is to use the constant flow rate until the
pressure builds up to a certain level and then to use constant pressure filtration
for the remainder of the time. The cycle can conveniently be achieved by using
centrifugal pumping.
Briquetting
Coal briquetting is one of the oldest applications in the forming process by
agglomeration. Originally, piston presses were used to produce coal bricks
of about 0.5 to 10 kg, made of coal fines that were mixed with binders such
as pitch, bitumen, or tar. These presses had reduced capacities (1,000 bricks
per hour maximum). The development of the double roll press (see Figure 2)
took place during the second half of the 19th century, and it offered higher
production capacities under acceptable economic conditions. Thanks to this
technique, the annual production has reached up to 10 million metric tons of
briquettes. Coal briquetting technology by means of the double roll press was
mainly developed for the upgrading of coal fines coming from coal screening
and washing, and used as fuel for domestic and/or industrial heating in the
same way as coal.
The new coal briquetting units are located in the larger coal-producing
countries (China and India), the developing countries (Turkey), or used for
coke production in Japan, India, and China by upgrading of lower-quality cok-
ing coals.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing Developments 101
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
102 Technology Developments and Plant Installations
S U M M A R Y O F A AT C F L O TAT I O N I N T H E W I T B A N K
C O A L F I E L D A N D D E V E L O P M E N T O F T H E M U LT I C E L L
Until the early 1990s, it was thought that Witbank coal was not amenable to
froth flotation. The first flotation plant to treat Witbank ultrafine coal was only
installed in 1995. Since that first plant, there have been more installations, par-
ticularly the 30-t/h Kleinkopje (KK) fines plant, utilizing a Multotec column
flotation cell and a Latham plate-and-frame filter press, which was constructed
and commissioned in 1998. The KK flotation plant was relatively successful
but was adversely impacted by weathered and burnt (spontaneous combustion)
feed material. In addition, the performance in the finer fraction (<100 µm) was
not acceptable, primarily as a function of bubble size. The Latham filter press
had many technical challenges but performed relatively well. Ownership of the
fines plant by plant personnel was a major challenge, as they viewed it as non-
core and this impacted negatively on the performance of this plant.
Generally, the Witbank flotation test work and practice steered away from
the traditional flotation technology routes as they were not considered suc-
cessful. It was highly accepted that high shear mixing was necessary to achieve
success. In addition, it was believed at the time that there was no justification
to float and produce steam coal. Steam coal generally attracts a low price, which
is made worse by the adjustment made for moisture. This caused metallurgists
to investigate other forms of dewatering inclusive of thermal drying, which is
often considered economically unjustifiable.
With new developments in flotation technology and flotation reagents,
the decision was made to conduct further flotation tests at Goedehoop col-
liery. Initial test work was conducted on a conventional Wemco mechanical
test cell using new, improved reagents, and promising yields were obtained.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing Developments 103
The Multicell
The Multicell incorporates a primary and secondary flotation cell together with
a single Multicell flotation pump. The pump consists of a modified spindle
pump (i.e., the barrel) and the shaft and intake are modified to achieve an
efficient and high-dispersion energy flotation cell. The material is fed into the
barrel together with air and reagents. The material is then dispersed into the cell
through a grated “window” at which point the bubbles start rising to the top
of the cell. The froth (i.e., the product) overflows the “lip” of the tank into a
launder, whereas the primary cell discard is pumped to the secondary cell. A
distribution manifold disperses the material into the tank volume for addi-
tional retention time to increase the process yield. As there is no pump, a much
calmer flotation process is experienced (Opperman et al. 2002). The Multicell
schematic is shown in Figure 3.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
104 Technology Developments and Plant Installations
Air
Feed
Slurry Reagent
Air
D E W AT E R I N G C O N U N D R U M
With all the excitement around the Multicell, there was still a large hurdle to
overcome with regard to dewatering of this very fine flotation concentrate. This
flotation concentrate consisted of 150 by 0 µm particles and at times as much
as 80% <45 µm. Steam or thermal coal is sold on an energy basis, and thus
moisture must be kept to a minimum. Initial test work centered on Humboldt
screen-bowl centrifuges by feeding a combined flotation concentrate and spirals
product in a predetermined ratio. Test work indicated that a 75% spiral/25%
flotation ratio achieved the best results. The coarser spiral fraction is required
to recover the ultrafine fraction because it acts as a filter bed. The recovery at
this optimum was in the order of 90% but dropped off dramatically as the flota-
tion material (ratio) was increased and the surface moisture of the final product
also increased significantly. As most of the loss of recovery occurred through
the screen section of the unit, it was therefore decided to test a solid-bowl cen-
trifuge by converting the existing units from screen-bowl to solid-bowl. This
increased the recovery of the ultrafine material but at higher surface moisture,
which also impacted on the throughput of the unit.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing Developments 105
plant was constructed and commissioned, and first coal was produced in 2004.
The Multicell performed satisfactorily but was constrained by the solid-bowl
performance with concentrate recoveries in the order of 50%. The unrecovered
portion was lost to the effluent and ultimately to tailings as this recirculating
load of ultrafines could not be accommodated.
Dewatering Developments
During this Multicell installation period at Greenside, Goedehoop continued
to conduct dewatering test work. In an effort to glean industry best practice,
a team set out to visit several operations within South Africa comprised of
various commodities. This was primarily to establish what was being used suc-
cessfully elsewhere. What came out of this exercise was that plate-and-frame
presses were the route to pursue. Historically, there had been a perception that
plate-and-frame presses were expensive, difficult to operate, complex, and being
a batch process. This initiated an in-depth investigation into plate-and-frame
presses. What transpired from this work was that there were more advantages to
plate-and-frame press filtration than disadvantages. The main advantages were
• Lower cake moistures,
• Dewatering that did not require the addition of coarser material as a
facilitator,
• Product that was easier to handle, and
• Virtually 100% recovery and no circulating ultrafines load.
Equipment selection is not the only factor that determines the amenabil-
ity to filtration and final product moisture when considering different process
streams and ore bodies. Important particle characteristics include
• Particle size distribution, and
• Clay content.
Test work investigating the effect of various parameters and ore character-
istics on the final product moisture as well as the flux (the rate of dewatering,
i.e., l/h/m2) is critical. Additionally, when considering plate-and-frame filter
press units, the filter press functionality (i.e., feeding, membrane squeeze, and
air blowing) can have various effects on the final product moisture.
A trade-off study was conducted, and the Lasta filter press (see Figure 4)
from Ishigaki in Japan was chosen for the envisaged Goedehoop fines plant
based on superior technical performance. In addition, the Lasta unit made use
of technology considered to be essential to ensure total cake discharge, namely,
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
106 Technology Developments and Plant Installations
traveling cloths and also automated cloth washing to ensure that no cloth blind-
ing takes place.
At this stage of the journey, considerable work was done on thermal dry-
ing, as it was believed that the moisture content in the filter cake may adversely
impact the overall gross as-received quality of the final product, expressed in
kcal/kg. Test work was successfully conducted on flash driers with acceptable
results. Any moisture could be achieved based on retention time in the unit.
It was generally believed that going below 8% was counterproductive as there
would be re-adsorption from the atmosphere and surrounding coal as the cake
would be added to the main product stream (the carrier) prior to stockpiling
and train loading.
This route was discontinued, however, as it was believed that the safety risk
in terms of fire and explosion was too great; the lack of expertise and the capital
requirement impacted on the feasibility of the project. In addition the calcu-
lated impact on the overall energy content of the total AATC product grade
complement was found to be small (approximately 10 kcal/kg).
G O E D E H O O P F I N E S P L A N T I N S TA L L AT I O N
Toward the end of 2005, Goedehoop submitted a capital application to con-
struct a fines beneficiation plant. The application was approved in October
2005, and commissioning of the new plant commenced by mid-2007. The
project was aimed at implementing a fine coal beneficiation plant making use
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing Developments 107
Positives
The average product filter cake surface moisture of 18.9% was achieved com-
pared to the target value of 20%. The filter cloth life exceeded the guaranteed
5,000 cycles.
Negatives
The flotation plant has not achieved the yield increase of 3%. This is due partly to
head grade issues but also inherent constraints with the Multicell design, namely:
• A throughput constraint as a result of the primary pump, which dou-
bles up as both a transfer mechanism and high-energy shearer, which
was further impacted by the conscious decision to lift the secondary
cells up for the tailings to gravitate to the thickener. This resulted in
lower throughput and higher feed solids.
• The low energy and recoveries in the secondary cells due to poor cell
dynamics.
The Goedehoop fines plant produced in excess of 15,000 tons per month, but
this did require a large amount of attention, technical input, and ownership by
plant management and operational teams.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
108 Technology Developments and Plant Installations
A G G L O M E R AT I O N
Anglo Thermal and Enprotec embarked on an ultrafine coal agglomeration
feasibility study in 2009. A decision was made to conduct the test work at a
specific operation where the ultrafine coal was being stockpiled after dewater-
ing the material using Technicas Hydraulicas plate-and-frame filter presses. This
material presented constraints as a result of the particle size with the power
utility refusing to accept this material.
Briquetting and granulation are two of the most widely used agglomera-
tion processes. Technology and affordable binders have made it possible for the
South African market to seriously consider briquetting to solve the ultrafine
coal challenge and, in the process, clean up the environment. Both briquetting
and granulation tests were conducted.
The aim of the project was to evaluate the influence of variables on the
final briquette and granule. This included the influence of varying binder addi-
tions, the effect of a sorbent, the optimum feed moisture, green strength, cured
strength, water proofing properties, and so forth.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing Developments 109
Briquetting Results
On completion of all the tests, it was evident that the trend with regard to bri-
quette strength is the same (i.e., as the binder dosage increases, the briquettes
can withstand more 2-m drops compared to the same briquettes at a lower
binder dosage). Another observation was that the waterproof characteristics of
the briquettes improve as the binder dosage increases.
Granules Results
Notably, there exists a feed moisture range where granulation is possible. When
the feed moisture drops below the range, no granulation will occur, as the feed
is too dry to ensure sufficient blending of the binder. Similarly, when the feed
moisture exceeds the range, a pulp will form rather than granules. With higher
binder addition and also higher feed moisture, the granules can withstand long
periods of time within water without decomposing. Importantly, the freshly
produced granule had no strength or waterproof properties. It also sticks to
other granules and forms large agglomerates. Fresh granules are not a final
product and need to be thermally dried before stockpiling. As a result of the
test work conducted, the decision was made to follow the briquetting route.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
110 Technology Developments and Plant Installations
Some cells had a shallow froth depth of 30 cm and there was a potential of
recovering a lot of gangue material. This was easily rectified.
The bubble size was found to be fine enough to float such fine material, as
shown by the Sauter mean bubble diameter of <0.5 mm. This confirmed the
advantage that the Multicell had over traditional flotation technology. This
indicated that there were challenges specifically around the throughput and sec-
ondary cell performance that appeared to be inherent to the Multicell design.
As time went by, the ownership deteriorated and skills were also lost as person-
nel were transferred, which resulted in declining performance. This initiated an
evaluation study utilizing a third party, namely Enprotec Mineral Processing.
E N P R O T E C P E R F O R M A N C E E VA L U AT I O N O F T H E
G O E D E H O O P F L O TAT I O N C I R C U I T A N D C O N V E R S I O N
TO DUAL CELL
Outcomes from this investigation confirmed previous findings. Challenges with
the Multicell system were due to the fact that the mounted pump is utilized for
both pumping as well as agitation. If, for whatever reason, the pump efficiency
and thus the pumping capacity is lost, the plant throughput is reduced and the
retention time of material within the cell is increased. An increase in retention
time increases the amount of discard material floated, effectively reducing the
efficiency of the process. A reduction in plant throughput causes an increase in
feed density due to an increasing recirculation load. The increase in feed density
overloads the froth phase (i.e., the froth carrying capacity is exceeded), which
reduces the yield.
The feed density was regulated by the dilution water addition and the
thickener underflow variable-speed drive (VSD) pump. Therefore, as soon as
the feed density to the plant is above the set point, the thickener underflow
VSD pump slows down, thus reducing the flow rate to the flotation feed tank.
The problem, however, is that the feed density was therefore approaching the
thickener underflow density (see Figure 5). This could be explained by the
fact that the minimum speed of the VSD pump is set at 50%. In addition, one
feed tank feeds one bank consisting of three process streams; therefore, to run
at optimal conditions, all three streams should be operational. It can thus be
shown that the feed density tends to the thickener underflow density when the
material is not moving through the cells fast enough as a result of the primary
pump transfer capability challenge as explained previously.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing Developments 111
Approaches Thickener
Underflow Density
Figure 5 Trends illustrating how the flotation feed density approaches the thickener
underflow density
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
112 Technology Developments and Plant Installations
Feed
Slurry Reagent - 75%
Air
Air
Tailings
Dual-Cell Not to Scale
Main Objectives
The main objectives were as follows:
• Increase the throughput. It must not be constrained by the primary
pump’s inability to pump slurry and air.
• Improve performance of the flotation plant by addressing dynamics in
the secondary cells.
The following changes were made such that the feed is gravitated through the
process system:
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing Developments 113
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
114 Technology Developments and Plant Installations
use of a single larger pump in the primary tank to transfer pulp to the
secondary tank.
These changes increased the throughput of the plant by removing the flow
constraint as well as increased the yield to the product stream at the specified
product quality by duplicating the primary in the secondary cell.
Pre- (Multicell) and postconversion (Dual Cell) performance measure-
ment tests were conducted on module B, and the results are summarized in
Table 1 (Kruger et al. 2011).
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing Developments 115
liability. Since commissioning, these yields were not achieved as a result of the
same Multicell inherent constraints experienced at Goedehoop.
Greenside colliery was thickening their flotation concentrate, blending it
with spiral product, and dewatering it using solid-bowl centrifuges (see Fig-
ure 8). This process had the following major drawbacks:
• High product moisture content
• Significant loss of product in the effluent
• Product that is difficult to handle
The high maintenance cost, low availability, low recovery, and high product
moistures of the three solid-bowl centrifuges resulted in a total yield loss of at
least 1.4% of the forecasted 3%.
Goedehoop colliery was making use of two fully automatic filter presses
to dewater their flotation concentrate. The problems that Greenside was facing
had been eliminated at Goedehoop due to the installation of filter presses. The
filters in use at Goedehoop had unfortunately become very expensive due to the
weakening of the rand. This fact was a key driver in Greenside deciding to test
other filter presses on the market that could achieve the following objectives:
• Fully automatic
• Low in capital and operational cost
• Produce acceptable final moisture content
• Produce a product that is easier to handle
• Operate at a high availability/utilization
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
116 Technology Developments and Plant Installations
Enprotec was at the time using and promoting a filter press from Sepro-
tech to dry their flotation concentrate, which had the characteristics required
by Greenside. The filter is known as the rapid filter press and is manufactured
locally in South Africa. Enprotec together with Anglo Thermal Greenside col-
liery embarked on a project to test the rapid filter press at Greenside colliery
in 2008 using an arising flotation concentrate as feed to the filter. The project
was divided into two phases with thickened flotation concentrate and flotation
concentrate used as feed in each respective phase.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing Developments 117
discovery of Jingjin and the presses they provided. The presses were custom
designed and cost effective, a further reduction of ±50% when compared to the
Rapid press, and they were of acceptable quality.
The major risk was that the Jingjin press (see Figure 9) would be a first in
South Africa. This risk was mitigated as follows. A technical visit was made to
the Jingjin factory in October 2010. The visit also included viewing the filters
operating in the field in coal applications. The Jingjin factory visit revealed a
professional, quality-controlled manufacturing process using computerized
machinery at all stages of manufacture. All critical components of the machine
are manufactured in-house. Quality control was of an acceptable standard and
the stock holding of all components was extensive.
F I LT E R P R E S S C O M PA R I S O N
A comparison between the Ishigaki, Seprotech, and Jingjin filter presses was
conducted and the Jingjin was chosen based on the following:
• Fit for purpose
• Cost effective
• User-friendly maintenance
• Operational understanding
The new fines plant at Greenside, which consists of four primary and
secondary cells converted to Dual Cell and two 1.5-m plate Jingjin presses, is
currently in the commissioning phase (see Figure 10).
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
118 Technology Developments and Plant Installations
I M P L E M E N TAT I O N O F B R I Q U E T T I N G AT A AT C
Dewatered fines are transported to a fines stockpile area for storage. Market-
ability of the fines in its current form is very limited. Because of environmental
concerns and the availability of technology, the decision was made to build
a 56-tph briquetting plant. The aim of the project was to take the stockpiled
fines and produce a sized salable product (refer to the conceptual layout of the
briquetting plant in Figure 11). The briquette plant’s main process streams are
briefly described as follows.
• Stream 1: This assumes optimum feed (surface) moisture of 15%–20%
to be loaded at 66 tph from a stockpile prepared by a front-end loader
(FEL). From the feed bin, material is fed onto a conveyor discharging
into the Eirich mixer.
• Stream 2: The binder addition takes place continuously in a range of
0.3%–0.5% by mass. The addition percentage is reassessed from time
to time. The dissolution of binder and the operation of the binder
plant are controlled by a programmable logic controller.
• Stream 3: After thorough mixing, the Eirich mixer discharges onto
a conveyor that feeds the briquette machine. The briquette machine
product is then transferred to the stacker conveyor for final stacking.
• Stream 4: Fines screening allow fines generated by the briquetting
process to be screened out and discharged onto a fines stockpile via
a fines stockpile conveyor. These fines are then circulated to the main
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing Developments 119
prepared feed stockpile via FEL ready to be loaded back into the
mixers.
• Stream 5: An Osborn KPI Series 33-30130 Super Stacker with luffing
and slewing capability places the briquettes on the stockpile area for
air drying to take place. Five days of rolling stock will allow the bri-
quette moisture to reduce and the binder to cure.
• Stream 6: Water is drawn from the clear water dam and is used for the
binder solution.
• Stream 7: Potable water is required for steam generation within the
boiler section of the binder plant.
• Stream 8: Binder is delivered by bulk truck and off-loaded into the
binder storage bin. Binder at the required percentage is added into the
mixing tanks and prepared for addition to the mixers.
The briquetting plant was commissioned and first briquettes produced in
January 2012. There have been challenges, most notably the following:
• High % fines creation post the briquetting machine—green strength
• Impact of the weather has been significant on briquette curing strength
and time taken to cure
• Inconsistent briquette strength
In essence, the cause of the challenges referred to can be summarized as follows:
• Inadequate and homogeneous binder mixing
• Briquette machine speed setting to ensure efficient pocket filling
• Transfer points and stacker luffing contributing to the fines creation
FUTURE PLANS
Flotation
Plans have been made to convert the remaining flotation streams at Goedehoop
from Multicell to Dual Cell during 2013 as a result of the improved recovery
and quality control experienced since conversion of the B Module was initiated.
The findings of the project concluded that the Dual Cell delivered improved
recoveries (yields) and improved quality control due to the fundamentally
improved flotation cell dynamics. An improvement of 0.4% yield (percentage
of total feed to plant) was experienced with a concurrent 0.34 MJ/kg improve-
ment in product quality. In addition, because of higher recoveries and increase
in froth solids content to filtration, optimum dewatering can be facilitated.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
120 Technology Developments and Plant Installations
CONCLUSION
South African coals, specifically the Witbank Coalfield, are classified as gener-
ally difficult to beneficiate. This holds true for fine coal, particularly the ultrafine
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing Developments 121
fraction that was considered not amenable to flotation until the mid-1990s and
also presents major difficulties in terms of dewatering. This has resulted in large
quantities of unbeneficiated ultrafine coal being stored in slimes compartments.
The journey of fine coal processing in AATC outlined in the chapter has
confirmed that Witbank ultrafine coal can be floated employing Multicell
flotation technology, which has evolved further in the Dual Cell. The ultrafine
flotation concentrate and thickener underflow can successfully be dewatered by
utilizing plate-and-frame filter presses, as proven by several installations within
AATC increasing salable thermal product and reducing the environmental
liability.
The briquetting of ultrafine material can be done successfully and econom-
ically, which will continue to add value, eliminate environmental liability, and
comply with South African regulations. The optimization journey in AATC
continues and will improve and evolve fine coal processing into the future.
ACKNOWLEDGMENTS
The author thanks his many colleagues at Anglo American Thermal Coal and
Enprotec Environmental and Process Technologies.
REFERENCES
Kruger, M., Vermaak, M.K.G., Buitendag, M.J., and Nxele, J. 2011. Coal flotation:
A case study of Dual Cell performance. South African Coal Processing Society
Conference.
Opperman, S.N., Nebbe, D., and Power, D. 2002. Flotation at Goedehoop Colliery.
South African Institute of Mining and Metallurgy. SA ISSN 0038-223X/3.00.
pp. 405–409.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing—
Technical Developments in
Australia
Bruce Firth, Mike O’Brien, and Graham O’Brien
ABSTRACT
Coal preparation plants incorporate a wide array of solid–solid and solid–liquid
separation equipment. The types of processes employed and configuration of these
unit operations vary considerably from plant to plant and across different geo-
graphic regions depending on the coal type and end use. Location-specific factors
influence plant design protocols and operating practices that need to be recognized.
This chapter highlights the differences between the Northern Hemisphere and
Australian requirements and includes some of the recent developments to overcome
problems arising in fine coal processing.
D I F F E R E N T N AT U R E O F R E G I O N A L C O A L S
There are both similarities and distinct differences in the design and operation
of Australian coal preparation plants with respect to other regions. This has
led to robust discussions and questioning of the respective approaches by these
parties. It does need to be recognized that for coal preparation, and many other
process industries, conservatism and fashion can dictate many decisions. But
there are some significant differences in the objectives of the respective regional
industries.
123
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
124 Technology Developments and Plant Installations
Markets
The markets for coal from the Northern Hemisphere and Australia are very
different. For example, more than 90% of U.S. production is for internal use in
thermal electricity production. There is a strict commercial association between
specific energy of the delivered coal and its value. Hence, both the ash value and
the moisture content of the coal are of importance. The Australian context is
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing—Some technical developments 125
strongly influenced by the international market for hard and soft coking coals
and pulverized fuel injection coals, with more than 80% of the coal production
exported. The ash value is of strict importance but the emphasis on moisture
is decreased with the recovery of material with prime coking characteristics
becoming an additional objective.
As a result, the majority of Australian plants operate at lower relative den-
sity of separations in their dense-medium cyclone (DMC) circuits compared
with the U.S. plants. The introduction of the pneumatic flotation cells ( Jame-
son and Microcel) with wash water applied in significant quantity to the thick
beds of froth has resulted in lower fine coal product ash values with a commen-
surate increase in the DMC separation density. This effectively has allowed the
plants to operate closer to the preferred situation having all cleaning circuits
operating at the same incremental ash value (Abbott 1981; Cierpisz and Gott-
fried 1977). If an intermediate cleaning step such as spirals is included, then it
would have difficulty in attaining a separation density of about 1.5, which is a
common value for the coking coals. This needs to be considered in conjunction
with potential problems of poor probable errors for the particles below 2 mm
in a DMC and flotation having difficulty with recovering plus 0.5-mm wedge
wire coal particles that may be present as the desliming screen wears.
In the United States, the density of separation is higher on average because
of favorable washabilities. The need to reduce the sulfur level to a reasonable
value is usually satisfied at these higher densities. The separation densities
among the dense-medium bath, DMCs, spirals, and flotation are much closer,
hence satisfying the optimization criteria of separating at the same incremental
ash value (Abbott 1981; Cierpisz and Gottfried 1977).
With regard to dewatering, the typical moisture content of Australian
export coal is in the order of 10% by mass for coking coal, whereas thermal coal
is mainly sold on a specific energy basis. The moisture level for the coking coal
is a result of a long-term compromise of recovering as much of the fine coal due
to its good coking properties at an acceptable moisture level. Vacuum filtration
has a long history of providing this need in Australia with horizontal belt filters
becoming the preferred unit due to ease of feed presentation and control.
Though there are a few installations of screen-bowl centrifuges in Australia,
this unit is dominant in the United States. It provides consistent low-moisture
values and performs a secondary size classification step by discarding signifi-
cant amounts of –0.04 mm material. Considerations have indicated that the
amount of water which this size fraction retains in the final product results
in it having a negative impact on the overall specific energy of a thermal coal
product (Hart et al. 2005; Luttrell et al. 2004). Another factor also needs to be
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
126 Technology Developments and Plant Installations
O V E R C O M I N G C L AY
Clay in run-of-mine coal is a major issue in the Australian coal preparation
industry. This is particularly evident in the Hunter Valley region in New South
Wales and the rapidly developing Surat Basin in Queensland. The majority of
the mines in these areas produce thermal coal products due to their lower rank
and/or the high ash value even after cleaning due to significant amounts of fine
clay embedded in the carbonaceous material.
The main fine coal processing issues arising from this type of raw coal are
• Removal of ultrafine material,
• The high amount of near-gravity material, and
• Continual breakdown of clay bands.
These issues give rise to a major problem with respect to the design of a coal
preparation plant for a new mine. The sample preparation approach to be used
for estimating the size distribution and washability of the raw coal has to be
carefully controlled, otherwise significant problems may arise with respect to
capacity in the fines circuit of a plant. There will be breakage due to handling
and continual breakdown of the clay/shale bands in the coal. This is critical to
the selection of process unit operations and their expected capacity require-
ments, particularly with regard to the fines circuit. Although there are some
variations on the approach in Australia, the work of Swanson et al. (1993) still
provides the basis.
Size Classification
Improvement in size classification at separation sizes below 0.25 mm has been
the subject of much investigation over a long period of time, and the classifi-
cation technology chosen for this separation is a compromise between many
factors such as
• Efficiency of the separation,
• Capacity of the unit operations,
• Level of operator attention and maintenance,
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing—Some technical developments 127
Hydrocyclone
A hydrocylone is the preferred unit because of its low cost, high capacity, and
simplicity in operation. The size of separation can be varied over a wide range
depending on the cyclone diameter and operating conditions. Unfortunately,
some significant deficiencies arise from this unit. These are clearly demon-
strated by data obtained by a detailed audit at a Hunter Valley coal preparation
plant. The fine coal circuit had two-stage desliming cyclones to remove the
ultrafine clay before the flotation circuit. There were two banks of 12 primary
cyclones and 12 secondary cyclones (all 250 mm in diameter).
Figures 1 and 2 show the size partition curves as well as those for selected
relative density fractions. Both cyclones had well-developed and stable spray
discharges. These results are typical of partition curves obtained from the other
cyclones.
The first finding is that for the individual particle density partition; there is
a significant difference in the separation sizes. The clays have a separation size of
about 0.02 mm, whereas the clean coal (relative density 1.28) is slightly above
0.10 mm and there is a major loss of coarse coal particles.
The ultrafine particles tend to follow the water split, and the fraction of
water reporting with the cyclone underflow (Rf ) is 0.07 and 0.20 for the pri-
mary and secondary cyclones, respectively. This is mainly due to the solids con-
tent of the respective feed flow streams being 10% and 18%, respectively. In the
Australian context, there are large amounts of clay in the ultrafine size fraction.
When the performance of the cyclones is considered with respect to par-
ticle size, it appears that a poor separation has been achieved. In Figure 2, the
size partition curve has a “bump” at about 0.05 mm size. This was observed
with some of the other cyclones and identified in some previous plant audits. It
was thought to be the result of poor laboratory analysis, but this investigation
has shown that it results from the density effect where the clay is separated at a
lower separation size than the coal and the washability of the feed coal.
Large-Diameter Hydrocyclone
The potential use of a 1-m-diameter classifying cyclone to separate fines in the
region of 0.2 to 0.3 mm would have some interesting advantages. There would
be reduction in capital and maintenance costs and a simpler plant layout. The
elimination of the slurry subdivision step would also lead to improved process
efficiencies. A detailed investigation of the size separation has been carried
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
128 Technology Developments and Plant Installations
0.9
0.8
0.7
Partition Coefficient
0.6
0.5
0.4
1.28
0.3
1.35
1.5
0.2 1.9
2.2
0.1 Size
Rf = 0.07
0
0 0.05 0.1 0.15 0.2 0.25 0.3
Size, mm
Figure 1 Individual partition curves for float/sink density fractions for one of the
primary cyclones
0.9
0.8
0.7
Partition Coefficient
0.6
0.5
0.4
1.28
0.3
1.35
1.5
0.2 1.9
2.2
0.1 Size
Rf = 0.20
0
0 0.1 0.2 0.3 0.4
Size, mm
Figure 2 Individual partition curves for float/sink density fractions for one of the
secondary cyclones
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing—Some technical developments 129
out at the Stratford coal preparation plant north of Newcastle in New South
Wales (O’Brien et al. 2000). A number of operating conditions, which could
be encountered in normal plant operation, were considered, including feed
solids content and vortex finder and spigot diameters. The separation efficiency
obtained for the cyclone was comparable to much smaller cyclones over a wide
range of conditions, and the separation size varied from about 0.15 to 0.35 mm
(see Figure 3).
Size by density float/sink analysis of the solids in the various flow streams
has shown that the shale has a separation size of about 0.08 mm, but the separa-
tion size of clean coal (float 1.3) can be varied from about 0.25 mm to nearly
0.5 mm (see Figure 4). This indicates that the solids content of the feed, its
washability, and subsequent cleaning unit operations need to be considered
carefully before a large unit of this type is installed in a plant.
The performance of the existing cyclone bank/rapped sieve bend circuit
was also determined for comparison (see Figure 5).
Sieve Bends
There have been many attempts to use sieve bends (rapped or unrapped) to
overcome some of the deficiencies of the hydrocyclone. Unfortunately, these
units tend to be low capacity, higher in cost, and require more operator and
maintenance attention. In fact, because of wear on the screen surface, it has to
be rotated at regular intervals. That is, the efficiency and capacity fall and rise
with the maintenance cycles.
The feed rate is the important variable with respect to the performance of
a sieve bend. There tends to be a feed rate at which a sieve bend will operate at
maximum efficiency. The efficiency declines at lower or higher values, and the
tolerance to variation increases for screens with larger apertures (Schreckengost
1989). The work of Slechta and Firth (1983) showed that this effect is due to
the efficiency of the undersize monotonically decreasing with increase in feed
rate and the efficiency of the oversize approaching a plateau. The capacity value
for the highest combined efficiency is the value at which the plateau is reached.
Figure 6 shows the impact of the mass flow rate over the sieve bend used by
Slechta and Firth (1983) on the performance of the sieve bend as described
by the partition curves. The partition curves start to degenerate for solids
flow rates above about 7 tph, which closely matches the 8.3 Lps/m value. The
capacity for maximum efficient separation has been estimated for a number of
apertures, and these are shown in Table 1. The solids content of the feeds were
all about 20% by mass.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
130 Technology Developments and Plant Installations
0.9
0.8
0.7
Partition Coefficient
0.6
0.5
0.4
0.3
0.2
10% solids
0.1
20% solids
0
0 0.2 0.4 0.6 0.8 1
Size, mm
0.9
0.8
0.7
Partition Coefficient
0.6
0.5
0.4
1.28
0.3
1.35
1.5
0.2 1.9
2.2
0.1 Size
Rf = 0.06
0
0 0.2 0.4 0.6
Size, mm
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing—Some technical developments 131
0.9
0.8
0.7
Partition Coefficient
0.6
0.5
0.4
0.3
0
0 0.2 0.4 0.6
Size, mm
Figure 5 Comparison between the size partition curves for the large-diameter
cyclone, the bank of conventional cyclones, and the complete cyclone bank/rapped
sieve bend circuit
Case Study
A recent paper (Bain 2012) provides an interesting case study of the changes
that a Hunter Valley mine had to make to overcome major clay problems, which
impacted on raw coal handling through to product quality. The plant was ini-
tially a DMC and spiral plant with no flotation. The mine has 19 seams with
different raw coal properties and washabilities. Nearly all of the potential plant
feed coals have high bentonite clay content, which has forced the plant circuit
to evolve over 20 years in the following manner:
• Low friction materials were required to facilitate flow of material
through the plant.
• Close monitoring of the amount of water being added at various parts
of the plant to prevent material “sticking” is required.
• The use of classifying cyclones only led to spiral products with high
clay contents, and hence high ash values and moistures. Initially, a sec-
ond bank of cyclones was added in series, which was aimed at remov-
ing all of the –0.08 mm material.
• The next step was to thicken the spiral product in a third bank of
cyclones prior to dewatering with a fine coal centrifuge.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
132 Technology Developments and Plant Installations
0.9
0.8
0.7
Partition Coefficient
0.6
0.5
0.4
0
0 0.05 0.1 0.15 0.2 0.25 0.3
Size, mm
Table 1 Estimated capacities for maximum separation efficiency for sieve bends
Screen Aperture, mm Capacity for Maximum Efficiency, Lps/m
0.120* 8.3
0.25† 16
0.3‡ 20
0.5 ‡
34
1.8‡ 55
3.5 ‡
103
* Rapped DSM sieve bend (Slechta and Firth 1983).
† Variesieve (Firth et al. 1995).
‡ Schreckengost 1989.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing—Some technical developments 133
concept of only having sieve bends for the classification was not viable
due to the amount of screen area required.
• Because the 0.125 mm separation size was not required, a 1-m-diam-
eter classifying cyclone was installed to replace the bank of smaller
cyclones.
• Jameson cells with the capability of significant addition of froth wash
water were employed for treating the ultrafine size fraction.
This circuit has resulted in product quality targets being met with major
increases in yield.
U S E O F C O A L G R A I N A N A LY S I S
The optimum design of a fine coal circuit requires a good understanding of the
associations of the minerals and macerals. This controls the washability of the
fines and the potential response to flotation. Coal grain analysis (CGA) char-
acterizes the individual fine coal grains in terms of liberated maceral and com-
posite particles (O’Brien et al. 2003), and hence enables an estimation of the
washability of the fines and the flotation response of specific grain types to be
determined. This new level of detail provides a potentially very useful approach
to process optimization and hence increased resource recovery.
CGA processes the individual coal grains separately and provides quantita-
tive data on particle area, which can then be used to estimate a volume composi-
tion (vitrinite, inertinite, dark mineral, and bright mineral). The density of the
grains can then be estimated from the densities of the individual components.
A summary of the information generated for a particular coal is shown in
Figure 7.
Figure 8 shows the washabilities of a feed coal to a spiral and the resulting
spiral product using CGA data. For comparison, conventional float/sink analy-
sis was also carried out and a reasonable match has been achieved. From these
data, partition curves for fine coal processing unit operations can be obtained
relatively quickly without the problems associated with organic liquid float/
sink analysis.
In a similar manner, the feed, concentrate, and tailings from a flotation
process can also be examined with respect to the recovery of the various
components in the feed. In this case, the components have been classified as
vitrite (greater than 95% of the grain is vitrinite), inertite, mineral, vitrinite
rich (vitrinite is the dominant maceral), inertinite rich, and mineral rich. The
results are shown in Figure 9 for the various size fractions. The vitrite is floating
relatively easily except for the coarse vitrinite-rich grains. Inertite is not floating
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
134 Technology Developments and Plant Installations
249,345 63,972 169,375 15,998 0 25.66 67.93 6.42 0.00 1.368 10.82
73,132 46,122 24,530 2,480 0 63.07 33.54 3.39 0.00 1.302 5.96
61,341 48,654 7,260 5,427 0 79.32 11.84 8.85 0.00 1.366 14.59
53,507 41,350 5,063 7,094 0 77.28 9.46 13.26 0.00 1.428 21.13
245,597 137,115 36,001 72,443 38 55.83 14.66 29.50 0.02 1.664 41.84
251,677 23,415 205,710 22,538 14 9.30 81.74 8.96 0.01 1.413 14.76
5,909 2,852 688 2,362 7 48.27 11.64 39.97 0.12 1.816 52.45
251,544 2,328 238,036 11,164 16 0.93 94.63 4.44 0.01 1.357 7.67
68,988 68,071 354 563 0 98.67 0.51 0.82 0.00 1.244 1.67
97,794 69,158 19,126 9,510 0 70.72 19.56 9.72 0.00 1.384 15.92
251,544 2,328 238,036 11,164 16 0.93 94.63 4.44 0.01 1.357 7.67
68,516 48,401 6,071 14,035 9 70.64 8.86 20.48 0.01 1.531 31.01
98,935 27,401 58,045 13,458 31 27.70 58.67 13.60 0.03 1.466 21.67
61,422 50,872 9,268 1,282 0 82.82 15.09 2.09 0.00 1.272 3.81
228,103 197,311 28,908 1,884 0 86.50 12.67 0.83 0.00 1.252 1.69
228,754 83,708 102,465 42,581 0 36.59 44.79 18.61 0.00 1.527 28.54
that well above 0.3 mm. Some of the locked mineral matter is reporting with
the product along with some ultrafine clay material. This technique allows coal
flotation to be considered in the same manner as used in mineral processing.
CONCLUSIONS
Development of improved fine coal processing circuits is progressing in Aus-
tralia within the constraints of the time and finance, and the coal quality issues.
New techniques to assist this development are becoming available and should
improve the recovery of more salable coal.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing—Some technical developments 135
100
90
80
70
Cumulative Recovery, %
60
50
40
30
20 CGA Feed
CGA Conc
Lab Assay Feed
10
Lab Assay Conc
0
1 1.5 2.0 2.5 3
Grain Density, g/cc
100.0
90.0
Min
Vit Rich C
60.0
In Rich C
Min Rich C
50.0
40.0
30.0
20.0
10.0
0.0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Characteristic Size, mm
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
136 Technology Developments and Plant Installations
ACKNOWLEDGMENTS
The authors thank CSIRO for permission to publish this information, ACARP
for jointly funding the research, and the many mines who have allowed access
to their plants so that the projects could be carried out.
REFERENCES
Abbott, J. 1981. The optimization of process parameters to maximize the profitability
from a three component blend. In Proceedings of the First Australian Coal Prepara-
tion Conference. Edited by A.R. Swanson. pp. 87–105.
Bain, G. 2012. Conquering the clay at Rix’s Creek CHPP. In Proceedings of the 14th
Australian Coal Preparation Conference. Edited by D. Mathewson. Paper D3.
Cierpisz, S., and Gottfried, B.S. 1977. Theoretical aspects of coal washer performance.
Int. J. Miner. Process. 4:261–278.
Cook, A.C. 1975. Australian Black Coal—Its Occurrence, Mining, Preparation and
Use. Symposium organized by the Illawarra Branch of the Australasian Institute of
Mining and Metallurgy. pp. 63–84.
Creech, M. 2002. Tuffaceous deposition in the Newcastle Coal Measures. Int. J. Coal
Geol. 51:185–214.
Firth, B., Edward, D., Clarkson, C., and O’Brien, M. 1995. The impact of fine classifi-
cation on coal preparation performance. In Proceedings of the 7th Australian Coal
Preparation Conference. Edited by J. Smitham. Paper E2.
Hart, G., Townsend, P., Morgan, G., and Firth, B. 2005. Improving fine coal centrifug-
ing—Stage 3. Australian Coal Association Research Program. Project C9047.
Luttrell, G., Honaker, R., and Yoon, R-H. 2004. Optimisation of the Coal Fuel Supply
Chain: A Coal Preparation Prospective. 29th International Technical Conference
on Coal Utilisation and Fuel Systems, Clearwater, FL. Preprint 140.
O’Brien, M., Firth, B., Clarkson, C., and Edward, D. 2000. Operational performance
of a large diameter (one metre) classifying cyclone. Proc. 8th Aust. Coal Prep. Conf.
C1: 98–111.
O’Brien, G., Jenkins, B., and Beath, H. 2003. Coal grain analysis for improved predic-
tion of utilisation performance. In Proceedings of the 12th International Conference
of Coal Science.
Schreckengost, D. 1989. Sieve bends handle high volumes. Coal 42–50.
Slechta, J., and Firth, B. 1983. Operation of a rapped sieve bend. Proc. Australas. Inst.
Min. Metall. 288:7–12.
Swanson, A., Fletcher, I., and Partridge, A. 1993. Improved Prediction of Size Distribu-
tions and Their Effects in Materials Handling and Coal Preparation Systems. Final
Report for NERDDC Project 1290.
van Krevelen, D.W. 1993. Coal: Typology, Physics, Chemistry, Constitution. New York:
Elsevier.
Whitmore, R.L. 1979. The washability of Australian coals. Proc. Australas. Inst. Min.
Metall. 270:47–53.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Beneficiation
Technologies 3
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing with
Dense-Medium Cyclones
G.J. de Korte
ABSTRACT
Dense-medium cyclones have been used for many years to beneficiate fine coal in
a number of countries around the world. The use of cyclones in this application is
not widespread, and at present, the process is employed only in South Africa and
China. The chapter provides a brief overview of past and current application of
dense-medium cyclones in the processing of fine coal and reviews some of the impor-
tant considerations for successful application of the technique.
INTRODUCTION
The dense-medium cyclone has, since its development by Dutch State Mines
(DSM) almost 70 years ago, become the main processing unit in the coal indus-
try. It is capable of efficient separations, even on coals containing high amounts
of near-density material, and accurate control of product quality is possible.
Although cyclones are capable of processing coal down to fine sizes, they are
generally applied only to coal coarser in size than 0.5 mm.
Beneficiation of fine coal, nominally below 0.5 mm, with dense-medium
cyclones was implemented in 1957, and cyclones are still being used to process
fine coal today, but history shows that the application of cyclones in fine coal
processing has not been nearly as successful as in the processing of coarser coal.
There is a perception in the coal processing industry that dense-medium
cyclone processing of fine coal is expensive, that it consumes much magnetite,
and that the separation efficiency is not good. There are, however, some who
realize that cyclones are still the most effective means to process fine coals
containing elevated amounts of near-density material at low cut-point densi-
ties, and, as such, the technology is presently in limited use in South Africa and
widely employed in China.
139
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
140 Beneficiation Technologies
BRIEF OVERVIEW
South Africa—1949
Some of the earliest work on beneficiation of fine coal with a dense-medium
cyclone was carried out in South Africa by the Fuel Research Institute (FRI)
of South Africa under the guidance of P.J. van der Walt in 1949. Using a
240-mm-diameter cyclone and barites as the dense medium, the results shown
in Table 1 were obtained when processing minus 1 mm by 0.25 mm coal. The
results showed that fine coal could be processed at a low relative density (RD)
and with very good separation efficiency.
Belgium—1957
In 1957, a flowsheet was developed by Stamicarbon/DSM and Evence Cop-
pée for a new dense-medium cyclone plant at Tertre in Belgium to process
10 × 0 mm raw coal. The plant was unique in that the feed to the plant was
not deslimed before processing. The raw coal was mixed with magnetite and
gravity-fed to two 500-mm-diameter (D) cyclones at a feed pressure of 9D.
The minus 0.75 mm coal in the feed drained through the product and discard
screens into the circulating medium tank. A part of the circulating medium
was pumped, using a separate pump, to two 350-mm-diameter cyclones at a
feed pressure of approximately 30D to affect a separation on the minus 0.75
mm coal. The overflow and underflow medium from the 350-mm-diameter
cyclones were directed to separate product and discard magnetic separators
to recover a fine coal product and discard in the underflows of the magnetic
separators. The product and discard were dewatered using conventional dewa-
tering equipment. A simplified flow diagram of the circuit employed at Tertre
is shown in Figure 1.
The separation efficiency of the plant at Tertre was good, and the results
were reported by Mengelers and Absil (1976) (see Table 2). The magnetite
consumption for the operation at Tertre was approximately 1 kg per feed ton.
In 1965, a similar plant was constructed at Winterslag in Belgium. This plant
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing with Dense-Medium Cyclones 141
10 × 0 mm
Raw Coal
10 × 0.75 mm
Discard
10 × 0.75 mm
Product
0.75 × 0 mm
0.75 × 0 mm Product
Discard
remained in operation for 17 years and was also reported to perform satisfacto-
rily (Lathioor and Osborne 1984).
United States—1970s
During the 1970s, a number of “wash-to-zero” plants, based on the Tertre
flowsheet, were built in the United States by Stamicarbon’s licensee, Roberts
and Schaefer. In these plants, efficient separation of the minus 0.5 mm size
fraction was not specifically targeted; therefore, the fine coal was processed,
together with the coarser coal, in a large-diameter cyclone. Recovery of the fine
coal from the drained medium was carried out in a similar manner as in the
Belgian plants. A simplified flowsheet for a typical wash-to-zero plant is shown
in Figure 2.
The results obtained from a wash-to-zero plant in West Kentucky, as
reported by Taylor and Chen (1980), are shown in Table 3. As can be seen, the
separation efficiency obtained on the fine coal was worse than that obtained
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
142 Beneficiation Technologies
50 × 0 mm
Raw Coal
50 × 1 mm 50 × 1 mm
1 × 0 mm Discard Product
Discard
1 × 0 mm
Product
in the Belgian plants. The finer coal was also separated at a much higher rela-
tive density than the coarser coal. Magnetite consumption for the plant was
reported to be around 0.5 kg/feed ton.
Similar results were reported by Island Creek Company, which con-
structed several wash-to-zero plants during the late 1970s and early 1980s.
Table 4 summarizes the results obtained at the Providence plant (Burch 1987).
Table 5 shows the names and commissioning dates for some of the wash-to-zero
plants built by Island Creek Company.
The main advantage offered by the wash-to-zero concept is simple plant
layout, which translates to low capital cost. The desliming screens and the fine
coal beneficiation section can be eliminated while still providing an efficient
separation on the coarse coal. The separation obtained on the fine coal, without
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing with Dense-Medium Cyclones 143
any additional equipment, is still better than that which can be had with most
water-based fine coal processing equipment.
Two plants designed to process 1 × 0.1 mm coal were also built in the
United States during the late 1970s. The plants were at Marrowbone in West
Virginia and Homer City in Pennsylvania. The coal preparation plant at Homer
City was built to prepare “deep cleaned” coal for burning in the power station.
The main purpose of the plant was to reduce the ash level and sulfur content
of the coal fed to the power station as an alternative to flue gas desulfurization.
The washability of the coal at Homer City is such that a very low cut-point
density of about 1.30 had to be employed to achieve coal of the required qual-
ity. Given that the minus 1 mm size fraction is beneficiated at a relatively higher
cut-point density than the coarser coal in a wash-to-zero operation and further
because this density cannot be controlled, it was necessary at Homer City to
employ a separate fine coal circuit to beneficiate the fine coal. The fine coal
plant was commissioned in July 1978. The performance of the plant during the
first year was reported to have been poor.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
144 Beneficiation Technologies
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing with Dense-Medium Cyclones 145
South Africa—1970s
In the early 1970s, South African coal companies concluded a contract with the
Japanese steel industry for the supply of a low ash blend coking coal. This coal
was to have an ash value of 7% and would be produced by processing the raw
coal from the Witbank no. 2 seam. The coal had to be processed at low rela-
tive density in the presence of very high amounts of near-density material, and
dense-medium separation was found to be the only process capable of affecting
the required separation. At the time, the fine coal could not be washed to 7%
ash and was therefore added raw to the lower-value middling coal.
Extensive research conducted in South Africa led to the conclusion that
dense-medium cyclones would be the only process capable of beneficiating the
fine coal to 7% ash (Horsfall 1976) and a 5-t/h pilot plant was subsequently
built at the FRI test facility in Pretoria in 1976. The plant was equipped
with a single 150-mm-diameter dense-medium cyclone supported by all the
additional equipment needed to feed the cyclone and recover magnetite. The
cyclone was fed at feed pressures typically ranging between 80 and 150 kPa
(40D to 70D). The plant was employed to test the beneficiation of fine coal
from several collieries in the country and proved that the production of low ash
coal from the fine fraction was possible. A simplified flow diagram of the FRI
fine coal dense-medium plant is shown in Figure 3.
The results obtained during a 10-day continuous operating period of the
pilot plant are shown in Table 6 (Fourie et al. 1980). Based on the success of
the pilot-plant work, a full-scale fine coal dense-medium plant was built at
Greenside colliery and commissioned in 1980. This plant produced low ash
coal from the fines fraction and remained in operation for 18 years. Three
additional plants, based on the FRI design, were constructed in South Africa
at Newcastle-Platberg colliery, High Carbon Products in Natal, and Rooiberg
tin mine. The 150-mm-diameter dense-medium cyclone used in the FRI pilot
plant is shown in Figure 4 (CSIR 1980).
Australia—1990s
A fine coal dense-medium plant was constructed at ARCO’s Curragh mine
in Central Queensland, Australia, in the early 1990s. As in South Africa, the
motivation for the plant at Curragh was to produce a product containing 7%
ash from the minus 0.5 mm size fraction.
Much of what had previously been learned about dense-medium cleaning
of fine coal was incorporated into the design of the Curragh plant, which was
carried out in cooperation with CLI Corporation of Pittsburgh, Pa. The lessons
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
146 Beneficiation Technologies
–100 micron
Product
Raw –1 mm
Discard
Figure 3 Simplified flow diagram of FRI dense-medium fine coal pilot plant
Table 6 Results obtained from FRI dense-medium fine coal pilot plant
Test No.
1 2 3 4
Product ash, % 6.5 7.7 7.4 5.9
Product yield, % 53.1 64.2 58.8 43.3
RD of separation 1.46 1.50 1.48 1.43
EPM 0.022 0.031 0.025 0.020
Total magnetite losses, kg/t 0.81 1.11 1.32 0.9
learned from the Homer City plant (United States) in particular were taken
into account (Kempnich et al. 1993).
The plant was equipped with two 500-mm-diameter dense-medium
cyclones. Much attention was devoted to the design of the magnetite-recovery
circuits, as well as to the desliming of the feed coal. Vibrating sieve bends were
used to assist in both desliming of the feed and recovery of medium. The plant
started up “smoothly” during commissioning. Magnetite consumption seemed
low and there was no noticeable impact on the magnetite consumption of the
overall Curragh operation. However, problems were encountered with the effi-
ciency of the cyclones, and it was not possible to produce the required 7% ash
level in the product. Test work carried out onsite led to the conclusion that the
initial cyclone feed pressure of 9D, or approximately 45 kPa, was too low. The
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing with Dense-Medium Cyclones 147
feed pressure was therefore increased to 120 kPa after which a 7% ash product
was consistently achieved (Kempnich et al. 1993).
The fine coal plant at Curragh closed down toward the end of 1996 for
reasons unknown.
South Africa—2001
In 2001, a 25-t/h fine coal dense-medium pilot plant was built in South Africa
by the Coaltech research program. The objective of the plant was to demon-
strate that coal from the Witbank no. 4 seam, which is difficult to process, can
be upgraded to export quality, something that could not be done using spirals.
The plant employed two dense-medium cyclones in series to improve separa-
tion efficiency and was run on standard “medium” grade magnetite—one of the
objectives for the research. A simplified flow diagram of the Coaltech circuit is
shown in Figure 5.
Extensive test work was conducted on the raw fine coal from several col-
lieries over a period of 5 years and much valuable information was gained from
the plant. As a result of the work done, the spiral plant at Leeuwpan colliery
was replaced with a dense-medium cyclone plant. This plant is presently in
operation at Leeuwpan. The dense-medium cyclones in the Leeuwpan plant
are shown in Figure 6, a photograph taken during the construction of the plant.
China—Recently
In recent years, many plants were built in China using the wash-to-zero
approach to beneficiate coarse and fine coal simultaneously without the use of
additional fine coal processing equipment.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
148 Beneficiation Technologies
–100 micron
Product
Raw –1 mm
Discard
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing with Dense-Medium Cyclones 149
beneficiated to yield a coking and a middling coal. The coal is recovered from
the small cyclone overflow and underflow with magnetic separators. The circuit
employed is similar to the one implemented at Tertre (Belgium) 55 years ago,
but by using the overflow medium from the three-product cyclones, fine mag-
netite is automatically ensured. The separation of the fine coal is reported to be
effective (Zhao and Yu 2012), and some typical results are shown in Table 7.
L E S S O N S L E A R N E D F R O M T H E PA S T
From the preceding overview on the subject of dense-medium processing of
fine coal, one can conclude that the use of dense medium to clean fine coal has
not always been easy or completely successful. Mixed results were obtained, but
some useful learning does emerge. Some of the more pertinent aspects of the
process are briefly discussed in this section.
Magnetite Sizing
Magnetite size consistency is perhaps the most confusing aspect of fine coal
processing. It is well known that finer magnetite improves the separation effi-
ciency in cyclones for the plus 0.5 mm size fractions (DSM 1970; de Korte
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
150 Beneficiation Technologies
2007). In general, it is assumed that for finer sized coal, even finer magnetite is
required, and the FRI (Fourie et al. 1980) insisted that magnetite which is 50%
finer than 10 μm is necessary for efficient separation in the 150-mm-diameter
cyclone. Sokaski and Geer (1963) found that the best separation could be
obtained using grade “B” magnetite. They further found that using finer magne-
tite improved the cyclone performance for coal particles coarser than 350 µm,
but no improvement was noted for the finer sizes.
Tests conducted in a 250-mm-diameter cyclone, operating at a feed pres-
sure of 117 kPa and using saturated brine as a medium (Mengelers 1982), indi-
cated that results which were very similar to those obtained using magnetite
were obtained. Table 8 shows a summary of the results obtained by Mengelers.
Notably, there is a differential of between 0.12 and 0.21 relating the
medium relative density of 1.20 and the actual relative density of separation.
The EPM values obtained are similar to those obtained elsewhere with magne-
tite. Figure 7 shows the normalized partition data for the brine tests compared
to the normalized results obtained at Homer City on a 200-mm-diameter
cyclone operating at a feed pressure of 70 kPa and using grade B magnetite
(Esposito and Higgens 1982). One can see that the results compare well.
The results obtained by Deurbrouck (1974) when processing fine coal
sized between 350 µm and 150 µm, using zinc chloride as the medium, also
indicated EPM values in the range of 0.04 to 0.065. These values compare well
to the results obtained when processing fine coal using magnetite as a medium.
It is difficult to conclude from these results what grade of magnetite would
be best in practice, but it would most probably be safe to assume that the fin-
est magnetite available should be used. Cyclone overflow medium from coarse
coal cyclones can be effectively used to provide fine magnetite to a fine coal
processing circuit. This is the practice at the Leeuwpan mine and also in the
Chinese plants.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing with Dense-Medium Cyclones 151
100
90 Homer City Tests
Brine Tests
80
70
Partition No.
60
50
40
30
20
10
0
0.85 0.95 1.05 1.15 1.25 1.35 1.45 1.55
RD/d50
45 kPa. When the plant started operation, the required 7% ash product could
not be produced. When the cyclone feed pressure was increased, initially to
65 kPa and then to 120 kPa, the required low ash product could be obtained
(Kempnich et al. 1993). The influence of the cyclone feed pressure on the per-
formance of the cyclones at Curragh is summarized in Table 9.
Tests conducted at the Coaltech pilot plant in South Africa indicated that
the normalized EPM value obtained when processing 1 × 0.1 mm coal could
be improved from 0.066 at 18D to 0.028 at 24D (McGonigal and de Korte
2005). Tests conducted at Homer City, however, found that better separation
efficiency was obtained at low cyclone feed pressures when processing fine coal
at very low relative density (Chedgy et al. 1986). It therefore seems that specific
conditions can dictate the required cyclone feed pressure.
Magnetite Consumption
The magnetite consumption reported from the majority of dense-medium
fine coal plants, both stand-alone and wash-to-zero plants, mostly fall within
normal limits. In general, these plants have magnetite consumption of between
1.0 and 1.5 kg/feed ton, which compares favorably with that of conventional
dense-medium plants processing deslimed feed. In the modern Chinese wash-
to-zero plants, magnetite consumption of around 0.55 kg/t is reported (Zhao
and Yu 2012).
In recent years, the performance of magnetic separators has improved
dramatically through the use of more sophisticated magnetic materials. The
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
152 Beneficiation Technologies
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing with Dense-Medium Cyclones 153
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
154 Beneficiation Technologies
900
800
700
600
500
Counts
400
300
200
100
0
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58
Size, micron
700
600
500
400
Counts
300
200
100
0
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58
Size, micron
Figure 10 Size distribution of magnetite lost via discard magnetic separator
material can be removed from the product after beneficiation. The fine coal
below 100 µm, being poorly separated and normally having an ash value higher
than that of the required final product, can degrade the quality of the final
product. Removal of the ultrafines from the product will therefore improve its
quality. This also applies in the case of other fine coal processes such as spirals
or teetered-bed separators.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing with Dense-Medium Cyclones 155
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
156 Beneficiation Technologies
1.60
1.55
1.50
RD
1.45
1.40
1.35
1.30
0.000 0.500 1.000 1.500 2.000 2.500
Particle Size, mm
0.12
0.10
0.08
EPM
0.06
0.04
0.02
0.00
0.000 0.500 1.000 1.500 2.000 2.500
Particle Size, mm
CONCLUSION
Dense-medium processing of fine coal is still the most efficient method of fine
coal cleaning available. It has been successfully implemented at several plants
around the world in the past and is presently in use in South Africa and China.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Processing with Dense-Medium Cyclones 157
REFERENCES
Burch, E. 1987. Island Creek’s feeding-to-zero concept simplifies coal prep circuit at
Providence plant. Min. Eng. 39:775–777.
Chedgy, D.G., Watters, L.A., and Higgens, S.T. 1986. Heavy medium cyclone separa-
tions at ultralow specific gravity. In 10th International Coal Preparation Congress,
Proceedings. Edmonton, Canada. pp. 60–79.
CSIR. 1980. Fuel Research of South Africa 1930–1980. CSIR Brochure B65.
de Korte, G.J. 2002. Dense-Medium Beneficiation of Fine Coal: Coaltech 2020 Pilot
Plant. CSIR Report no. 2002-0189. March.
de Korte, G.J. 2007. Dense medium cyclone tests conducted with coarse magnetite.
Presented at South Africa Coal Processing Society International Coal Conference:
Coal Processing in a Changing World, Johannesburg, South Africa, September.
Session 4, Paper 2.
Deurbrouck, A.W. 1974. Washing Fine-Size Coal in a Dense-Medium Cyclone. USBM
Report of Investigations 7982.
DSM (Dutch State Mines). 1970. The Heavy Medium Cyclone Washery for Minerals
and Coal. DSM.
Esposito, N.T., and Higgens, S.T. 1982. Modification of the fine coal circuit at Homer
City coal preparation plant. Paper presented at the SME-AIME Annual Meeting,
Dallas, TX, February 14–18.
Fourie, P.J.F., van der Walt, P.J., and Falcon, L.M. 1980. The beneficiation of minus
0.5 mm coal by dense medium cyclone. J. S. Afr. Inst. Min. Metall. (October):
357–361.
Horsfall, D.W. 1976. The treatment of fine coal: Upgrading –0.5 mm coal to obtain a
low-ash product. ChemSA ( July): 124–129.
Kempnich, R.J., van Barneveld, S., and Lusan, A. 1993. Dense medium cyclones on
fine coal—the Australian experience. In 6th Australian Coal Preparation Congress,
Proceedings. Paper E1.
King, R.P., and Juckes, A.H. 1986. Beneficiation of Fine Coal in a Dense Medium
Cyclone: The Effect of Slimes in the Feed. Report CSP Coal-6. University of the
Witwatersrand, Johannesburg, South Africa, January.
Lathioor, R.A., and Osborne, D.G. 1984. Dense medium cyclone cleaning of fine coal.
In 2nd International Conference on Hydrocyclones, Proceedings. BHRA The Fluid
Engineering Centre, Bath, England. Paper G1.
McGonigal, S., and de Korte, G.J. 2005. Dense-Medium Beneficiation of Fine Coal:
Coaltech 2020 Pilot Plant at Navigation, Landau Colliery. CSIR Report
2005-0566.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
158 Beneficiation Technologies
Mengelers, J. 1982. Report on Tests Carried Out at Stamicarbon with Coal 0–1.0 mm,
Treated in a Heavy Medium Cyclone of Diameter 250 mm. Stamicarbon BV Report
4950/Stac/95-1. March.
Mengelers, J., and Absil, J.H. 1976. Cleaning coal to zero in heavy medium cyclones.
Coal Min. Process. (May): 62–64.
Sehgal, R., Matoney, J.P., and Esposito, N.T. 1982. Innovative heavy media fine coal
cleaning: Homer City coal cleaning plant. In Ninth International Coal Preparation
Congress, Proceedings. New Delhi, India. Paper 143. C1–C15.
Sokaski, M., and Geer, M.R. 1963. Cleaning Unsized Fine Coal in a Dense-Medium
Cyclone plant. USBM Report of Investigations 6274.
Taylor, B.S., and Chen, W.L. 1980. Heavy medium cyclone cleaning of 1½" × 0 raw
coal. Paper presented at the 1980 American Mining Congress International Coal
Show, Chicago, IL, May 5–8.
Van der Walt, P.J. 1949. Operating characteristics of the cyclone washer. In Fuel
Research of South Africa Report No. 17.
Zhao, S., and Yu, J. 2012. Novel efficient and simplified coal preparation process. Inter-
national Coal Prep 2012, Lexington, KY, April 30–May 3. Paper 9.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of the
Reflux Classifier
K.P. Galvin
ABSTRACT
The reflux classifier is a new gravity separation technology for processing fine coal
and other dense minerals, developed through collaboration between the University
of Newcastle and Ludowici Australia. This chapter provides an account of that
development, covering a number of the key findings that have led to the final
design, and indeed the separation performance achieved. The chapter also outlines
the theoretical basis used to quantify the capacity of the separator. The technology
embodies the benefits of a high shear rate similar to that observed for spirals, the
benefits of an autogenous dense medium similar to that observed in a teetered-bed
separator, and an increased throughput advantage similar to that produced by
a lamellae thickener, with a strong synergy achieved among all three. Moreover,
the laminar shear mechanism that arises within closely spaced inclined channels
permits a powerful separation on the basis of density, delivering a significant per-
formance advantage at the low cut points required for the recovery of metallurgical
coal.
INTRODUCTION
The reflux classifier, shown in Figure 1, consists of a fluidized bed, with a sys-
tem of parallel inclined channels above. Feed enters the system and distributes
into an overflow and underflow stream in accordance with a mechanism that
amplifies the role of particle density by suppressing the effects of particle size.
Fluidization through the base is used to establish a suspension of higher-density
particles that acts as a dense medium. Lower-density particles thus transport
up through the system of inclined channels. Fine, dense particles entrained
into the inclined channels segregate from the flow and onto the inclined sur-
faces, joining the lower bed. When the density of the lower bed exceeds the set
point value, an underflow valve opens, discharging the high-density particles.
159
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
160 Beneficiation Technologies
Feed
Overflow
Fluidization
Underflow
Meanwhile. the lower-density particles emerge with the overflow. The physics
that underpins this separation is outlined in this chapter.
The reflux classifier, which followed a period of more than 10 years of col-
laborative research and development by the University of Newcastle (Australia)
and Ludowici Australia, is delivering significant benefits to the coal industry,
especially in the beneficiation of fine metallurgical coal. Similar performance
has been evident in the beneficiation of dense minerals (e.g., chromate, iron
ore, and mineral sands). This chapter provides a unique historical perspec-
tive behind the development of the technology, the rationale for the physical
arrangement, and the fundamental theoretical basis that governs the particle
separation. This chapter also outlines for the first time the theoretical basis used
to assess potential applications and in turn to predict solids throughput.
Consider the batch settling of fine particles suspended in a long tube. The
upper interface is observed to move downward at a steady rate. A slight tilt of
the tube, however, leads to a dramatic increase in the rate of collection of par-
ticles at the base of the vessel, with particles depositing first onto the inclined
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of the Reflux Classifier 161
surface, leading to the rapid sliding of the sediment toward the base, coupled
with the upward movement of the return fluid. Boycott (1920) was the first to
report about the effect of inclining a suspension, whereas Ponder (1925) and
Nakamura and Kuroda (1937) provided the first theoretical explanation, the
so-called PNK kinematic model. Other work led to the development of the
first industrial application, the lamellae thickener (see for example, Yao 1973).
Others (Carty 1957; Acrivos and Herbolzheimer 1979; Leighton and Acrivos
1986; MacTaggart et al. 1988; Davis et al. 1989; Zhang and Davis 1990; Smart
et al. 1993; Galvin et al. 2001) exploited the physical arrangement to explore
additional applications and some truly fundamental questions concerned with
the way particles interact with solid surfaces.
My first observation of the remarkable effects of inclined settling was
made during a visit to a laboratory in Houston, Texas, in 1992, concerned with
directional drilling in the oil industry. Later, a simple experimental system was
established consisting of a 25-mm-diameter tube, 1 m long inclined at 45° to
the horizontal, with feed entering near the base, underflow withdrawn from
the sediment near the base, and the balance of the flow reporting through the
top of the device as overflow (Thompson and Galvin 1997). The initial work
was concerned with the classification of fine coal on the basis of particle size,
whereas later work examined the potential to separate on density. The scale-up
of this single tube arrangement to meet the requirements of an industrial sepa-
rator were unclear.
At this time, teetered-bed separators (TBSs) were being introduced to the
Australian coal industry. The introduction of this technology was motivated
by the increasing need to overcome the limitations of spirals, that is, the need
to lower the separation density, and in turn recover the more valuable metal-
lurgical coal. I subsequently became involved in a study on TBSs, led by Stuart
Nicol. This work (Galvin et al. 1999) was funded by the Australian Coal
Association Research Program (ACARP), with direct support from Rio Tinto.
The benefit of an autogenous dense medium in promoting relatively large low-
density particles to segregate upward from the bed was in evidence, while the
considerable entrainment of fine dense particles into the overflow when the
feed pulp density was reduced was also apparent.
Following the completion of these two projects, there was a need to con-
ceive of a method to exploit the benefits of inclined settling. The problem
identified in scale-up of the inclined tube separator was solved by combining
the TBS arrangement with a system of parallel inclined channels. Though simi-
lar to a lamellae separator, this system incorporated the delivery of fluidization
water through the base of the vessel, allowing a fluidized bed to be established,
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
162 Beneficiation Technologies
with a mixing zone permitted to form in the zone between the top of the bed
and below the inclined plates. At this time, the arrangement was generalized to
cover multiple sets of inclined channels separated by mixing zones, as described
in the patent, a reflux classifier (Galvin 2001).
Intuitively, this basic arrangement made a great deal of sense, with a sig-
nificant increase in the effective sedimentation area generated by the presence
of the inclined plates. There should also be significant interaction between the
sediment sliding down the inclined surfaces and the upward fluidization of the
suspension from below.
RESEARCH PROGRAM
A Research and Development Agreement was established between the Uni-
versity of Newcastle and Ludowici Australia to develop this new technology,
referred to as a reflux classifier. This collaboration first developed through
Maurie Munro of Ludowici, and later through Taavi Orupold of Ludowici. The
research consisted of two main parts, a pilot-and-full-scale study at the nearby
Bloomfield coal preparation plant and work of a more fundamental nature at
the University of Newcastle. Funding was secured from ACARP and the Aus-
tralian Research Council.
Pilot-Scale Research
There was confidence that the pilot-scale study would lead to early success,
given the simplicity of the concept, coupled with our past experience and
hence understanding of the TBS. However, the commencement of the pilot-
scale research proved to be premature, with a series of design flaws evident
early on, requiring the reengineering of the whole system. Many problems were
encountered during the first 6 months of the study with the control of the feed
delivery, feed entry, and PID (proportional–integral–derivative) control of the
underflow discharge. The feed was also highly variable with high clay content
during the early period.
The selection of an inclined channel angle of θ = 60°, spacing of z = 60 mm,
and length of L = 1,000 mm was based on the objective of delivering a signifi-
cant solids throughput advantage over the TBS, while ensuring that the system
internal design was free-flowing and economic. The PNK model predicted a
throughput advantage of order Lcosθ/z~10. It was evident from the research,
however, that the throughput advantage was in the range of three- to fourfold.
The final work was published in Minerals Engineering (Galvin et al. 2002),
with partition curves presented for narrow particle size fractions over the range
0.25 to 2.0 mm. The composite probable error (Ep) reported for this overall
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of the Reflux Classifier 163
size range was 0.14. With a solids throughput of 47 t/m2h, the work, in the
end, proved to be sufficiently successful for a larger-scale trial to be undertaken.
U = U t (1 − φ )
n
(EQ 1)
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
164 Beneficiation Technologies
0.4
6.0 kg
Solids Volume Fraction
4.0 kg
0.2
2.0 kg
0.1
Conventional Fluidization
Expansion
0.0
0.0 0.01 0.02 ut 0.03 0.04
Fluidization Velocity, m/s
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of the Reflux Classifier 165
Figure 3 Original installation of the first RC1800 at the Bloomfield plant in 2003
and transport of the overflow coal was immediate, with uniform flow achieved
through each of the channels.
With the introduction of the new launder design, the channel spacing was
in fact increased to 120 mm by removing every second plate. This decision was
based on the discrepancy observed in the earlier pilot-scale work where the
throughput advantage of 3.5 was much lower than the kinematic theoretical
value of 10. This result suggested that half the inclined plates were redundant.
The available feed, at this time, was reduced to only 60 t/h; hence, it was not
possible to properly assess the effect of this change. However, several years later,
the additional plates were returned to the unit.
Other major hurdles were encountered during the full-scale trial work,
notably the effects of wear on the fluidization nozzles and significant effects
due to wear on the underflow valve. There were also problems encountered due
to the presence of grit in the fluidization water, which on occasions caused the
blockage of some of the nozzles, most notably those just below the pressure
transducers. The bed would lose fluidization in this location, which would
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
166 Beneficiation Technologies
undermine the control of the whole system. A novel device for removing this
grit solved the problem.
Finally, in late 2004, all of these problems were solved, and the plant pur-
chased the RC1800 and incorporated the device into their circuit on a perma-
nent basis. The composite Ep reported for the overall size range from 0.25 to
2.0 mm was 0.15 (Galvin et al. 2005). The unit continues to operate to this day.
Re 2s
> 32 (EQ 2)
Retn
where Res is the shear Reynolds number and Retn is the sedimentation Reynolds
number in the normal direction of the plane. This result has been expressed
using Reynolds numbers based on the particle diameter and, hence, has been
altered accordingly from the original form. Early work, which typically involved
inclined channels with a spacing of z = 60 mm, demonstrated no evidence or
even remote prospect for inertial lift. In fact, it was concluded that the inclined
channel spacing would need to be in the vicinity of a few millimeters and,
hence, concluded that the lift force criterion had little relevance to this work.
A comprehensive investigation of semi-batch elutriation from the new
laboratory system commenced, focused on the effect of the inclined channel
spacing. The PNK model had failed to explain the throughput advantage in the
earlier pilot-scale study. In this new work, a given system of particles was elutri-
ated at a specific superficial velocity. The number of equally spaced plates was
increased and the separation size determined. The separation size was found to
decrease. The number of equally spaced plates was increased again, producing
a further reduction in separation size. Eventually the separation size reached
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of the Reflux Classifier 167
1
η= (EQ 3)
1 + 0.133cos θ Re0.33
t ( L / z)
where θ is the angle of inclination with the horizontal, Ret is the particle Reyn-
olds number, and L/z is the channel aspect ratio. The term segregation efficiency
should not be aligned with the efficiency of the separation. Low segregation
efficiency implies that the particle has a strong tendency to convey, whereas,
conversely, higher segregation efficiency is consistent with a tendency to seg-
regate onto the inclined surface and slide downward. In fact, lower-density
particles tend to have lower segregation efficiency, which in turn promotes their
separation from the higher-density particles.
Using the approach of Ponder (1925) and Nakumara and Kuroda (1937),
the effective vessel area increases by the factor
L
F = 1 + cos θ sin θ (EQ 4)
z
U / U t = ηF (EQ 5)
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
168 Beneficiation Technologies
(2006) or lead to the inertial lift that had been sought since 2002. The results
were extraordinary, with a constant partition number obtained for particles
of different size ranging from about 0.1 to 0.4 mm. The partition curve, based
on particle size, was horizontal. Moreover, the result was repeated for particles
of different density, with a very different superficial velocity required in each
case (Galvin et al. 2009). Continuous steady-state experiments were then con-
ducted on metallurgical coal feeds (Galvin et al. 2010) using a wider channel of
z = 4.2 mm, with the composite Ep = 0.06 over the particle size range 0.25 to
2.0 mm. It is emphasized that this work was conducted at a relatively low solids
throughput of about 12 t/m2·h, with satisfactory performance achieved down
to a lower particle size of 0.075 mm. Data were also presented at significantly
higher solids throughput, with satisfactory performance maintained down to a
particle size of 0.25 mm. Figure 4 shows the results obtained at two cut points
and a comparison with the earlier work (Galvin et al. 2004). Figure 4 shows
the variation in the relative density of separation (d50) with particle size was
significantly lower than previously possible. Hence, there is a clear step change
in separation performance. Figure 5 shows that Ep values remained remarkably
low across a broad particle size range. The Ep values are much lower across the
size range than previously reported (Galvin et al. 2004). There was a clear step
change of improvement in separation performance.
When closely spaced inclined channels are used, the flow Reynolds number
decreases, resulting in the formation of a laminar flow profile, with a high shear
rate near the planar surfaces. The laminar flow profile produces a strong, and
very precise, variation in the local fluid velocity with the normal distance from
the wall. Thus, relatively fine particles, which migrate toward the upward fac-
ing planar surfaces during their transport through the channels, are ultimately
exposed to relatively low fluid velocities, while coarser particles are exposed to
proportionally larger fluid velocities. The velocity gradient at the wall delivers a
fluidization velocity to the particles that is proportional to the diameter of the
particle. This in turn reduces the effective dependence of the particle settling
velocity on the particle diameter by a full decade. Applied to particles that settle
within the intermediate settling regime, this mechanism effectively removes the
dependence of settling velocity on particle size. Fine, dense particles that reside
on the upward-facing inclined surfaces are exposed to low local fluid velocities
and therefore slide downward. Fine low-density particles are, of course, readily
conveyed, while much larger low-density particles experience a strong inertial
lift force that exposes such particles to even higher local fluid velocities, causing
them to lift and convey.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of the Reflux Classifier 169
2.4
2.0
Relative Density, d50
1.8
1.6
1.4
1.2
0.1 1.0 10.0
Particle Size, mm
Figure 4 d50 vs. particle size at two cut points as reported by Galvin et al. (2010) and
compared to that previously reported (Galvin et al. 2004)
0.05
0.00
0.1 1.0 10.0
Particle Size, mm
Figure 5 Ep vs. particle size at two cut points as reported by Galvin et al. (2010) and
compared to that previously reported (Galvin et al. 2004)
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
170 Beneficiation Technologies
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of the Reflux Classifier 171
Conversely, Figure 7 shows that for an Ep of 0.06, the yield of the lowest
density particles is nearly 100%; hence, a higher cut point of 1.53 can be used.
The target product, with an average relative density of 1.4, is then achieved
at a much higher yield of 30%. Figure 8 shows the improved performance in
terms of the actual yield relative to the maximum possible yield (for a perfect
separator) at an Ep of 0.06. Evidently, a product relative density as low as 1.3
can be obtained. Although a feed washability with an even density distribution
is extreme, this distribution is relevant to more poorly liberated coals as found
in India. With more typical coals, the benefits remain significant with a much
higher yield achieved via a low Ep.
Galvin and Liu (2011) developed a theoretical elutriation model of the
laminar-shear mechanism incorporating the inertial lift force model of King
and Leighton (1997). Under laminar flow, the particles at the upward-facing
surface of the inclined channel experience a local fluid velocity much lower
than the superficial velocity. The throughput advantage, based on this mecha-
nism, is approximately
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
172 Beneficiation Technologies
1.00
0.75
Partition Number
D50 1.53
Ep 0.06
0.50
RD product 1.40
Y 30%
0.25
D50 1.3
Ep 0.14
RD product 1.40
Y 12%
0.00
1.0 1.2 1.4 1.6 1.8 2.0 2.2
Relative Density
100
Ep = 0.06
75
Ep = 0.14
Relative Yield, %
50
25
0
1.2 1.3 1.4 1.5 1.6 1.7 1.8
Relative Density of Product
Figure 8 Relative yield vs. relative density of the product for Ep values of 0.06 and 0.14
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of the Reflux Classifier 173
U z
= (EQ 6)
U t 3d
where d is the diameter of the target particle and Ut is the corresponding termi-
nal velocity. Equation 6 applies to relatively fine particles that fail to experience
inertial lift. For larger particles, the shear rate leads to the lift force as
Figure 9 shows the theoretical elutriation achieved via the laminar shear
mechanism (Galvin and Liu 2011) for the original channel spacing of 1.77 mm,
in particular the strong potential to separate particles on the basis of density.
The original paper includes experimental data and hence reports the full valida-
tion of the theory, requiring no adjustable parameters. Figure 10 shows the lift
force fraction versus the channel spacing for a particle of density 2,600 kg/m3
and diameter of 0.310 mm in water subject to a superficial velocity of 0.05 and
0.1 m/s. The lift force fraction is the ratio of the buoyant weight of the particle
in the normal direction to the planar surface relative to the shear-induced lift
force. Clearly, the lift force fraction is negligible when the channel spacing is
large and does not reach a value of 1.0 until the channel spacing is relatively
small. It is evident that closely spaced channels are essential for producing suf-
ficient lift.
A P P L I C AT I O N O F T H E O R Y
This section is used to illustrate the application of the theory to the design of
the reflux classifier.
Design Calculations
The commercial reflux classifier is the RC2020, which has a diameter of 2.0 m
and, hence, cylindrical cross-sectional area of 3.14 m2. The solids feed flux is
always quoted relative to the value of this lower vessel area. The upper lamellae
section is constructed across a square section 2.0 m × 2.0 m = 4.0 m2, with an
overflow launder external to this section. The square system consists of rela-
tively large internal launders and 18 generic boxes of inclined plates, each of
internal horizontal dimensions of 0.653 m × 0.212 m = 0.138 m2. The inclined
plates are 0.212 m wide and 1.5 mm thick. In total, the horizontal cross-
sectional area of the boxes is 18 × 0.138 m2 = 2.49 m2; hence, the combined
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
174 Beneficiation Technologies
0.5
1,400 kgm–3 2,600 kgm–3 4,450 kgm–3
0.4
Particle Diameter, mm
0.3
0.2
0.1
0.0
0.00 0.02 0.04 0.06 0.08 0.10 0.12
Superficial Channel Velocity, m/s
1.0000
0.1000
0.1 m/s
Lift Force Fraction
0.0010
0.05 m/s
0.0001
0.1 1.0 10.0 100.0
Channel Spacing, mm
Figure 10 Lift force fraction vs. channel spacing for a particle of density 2,600 kg/m3
and diameter of 0.310 mm, subject to a superficial velocity of 0.05 and 0.1 m/s
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of the Reflux Classifier 175
area of the internal launders and internal walls is 4.0 – 2.49 = 1.51 m2. Thus,
the active box area relative to the area of the lower cylindrical cross section is
2.49/3.14 = 0.79.
The inclined channels are 1 m long, inclined at 70° to the horizontal, with
a 6-mm perpendicular spacing. For the generic application where the objective
is to achieve gravity separation over the particle diameter range 2 mm × 0.25
mm, the units are rated at 100 t/h of solids nominal and 120 t/h maximum.
The capacity of the unit, however, depends on the nature of the feed. For
example, a feed with very low density coal and little mid-range density mate-
rial can be processed at much higher rates than a feed with significant mid-
range density material. Thus, the so-called maximum rating may well be easily
achieved for a feed with little mid-range density material. But, of course, the
industry seeks nominal values, so in summary, a value of 100 t/h is stated for
this application.
Another key design factor is the yield, which dictates the relative solid
flows to overflow and underflow. At a low yield, relatively little material reports
to the overflow, and hence in principle the system is not constrained by the
capacity of the inclined channels. This suggests that the system could operate at
a higher feed solids throughput. However, this is a false assumption. The higher
solids rate to underflow results in a higher water split to underflow and, hence,
increasingly higher portions of particles become misplaced to the underflow.
Thus, a constraint is imposed requiring the water split to the underflow to be
10% or less. This constraint in turn limits the feed solids throughput to realistic
levels, and imposes the need to consider lower feed pulp density. In the follow-
ing section, this result is achieved by setting a lower value for the volume frac-
tion of solids in the overflow.
The design capacity is governed by the smallest dense particle to be
retained below the system of inclined channels. This particle is required to
ultimately report to the underflow. In the example, the target is a 0.25-mm-
diameter particle of density 1,900 kg/m3. The upper particle size requires more
judgment or direct knowledge of past performance, with a maximum diam-
eter ratio of eight times the lower target diameter considered appropriate. A
more conservative limit of fourfold could be equally applied. So the top size is
approximately 2 mm. The autogenous dense medium needs to prevent the low
density 2-mm-diameter particles from reporting to the underflow. By ensuring
that the right hydraulic flows into the inclined channels, there is also a strong
tendency for large low-density particles to convey to the overflow. The theoreti-
cal work associated with the laminar shear mechanism is very much concerned
with the mechanism that underpins this hydraulic conveying. However, when
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
176 Beneficiation Technologies
assessing the capacity of the reflux classifier in a given application, the primary
consideration is the smallest dense particles to be retained below the system
of inclined channels. This assessment, as described in the following section, is
based on Equations 3, 4, and 5.
Example Calculations
Consider an upper section of parallel inclined channels covering an open box
area of 0.653 m × 0.212 m. Each plate has a thickness of 1.5 mm, width of
0.212 m, active length to the overflow weir of 1.0 m, and angle of inclination
to the horizontal of 70°. The target particle has a diameter of 0.25 mm and a
density of 1,900 kg/m3. The design is based around preventing this particle
from reaching the overflow. This condition is established using the theory of
Laskovski et al. (2006). The choice of the density of 1,900 kg/m3 is based on
the anticipated D50 of the partition curve applicable to the target particle size.
The terminal velocity of the target particle of size d = 0.25 mm and density
1,900 kg/m3 settling in water is obtained using the Zigrang and Sylvester equa-
tion (1981) given by
2
d 1.5
0.5
(
Ret 14.51 + g ( ρs − ρ) ρ )
0.5
1.83 − 3.81 (EQ 8)
µ
where Ret is the particle Reynolds number, d is the particle diameter, ρs is the
particle density, ρ = 1,000 kg/m3 is the fluid density, μ = 0.001 Ns/m2 is the
fluid viscosity, and g = 9.8 m/s2 is the acceleration due to gravity. The particle
Reynolds number is
2
0.5
( (
Ret = 14.51 + 9.8 1,900 − 1,000 1,000 ) ) 1.83 × 0.000251.5 / 0.001
0.5
− 3.81 = 4.79
ρU t d
By definition, Ret = , and hence the terminal velocity is
µ
For 82 channels across the box width of 0.653 m, with plate thickness
1.5 mm, the perpendicular channel spacing, z = 6.0 mm. That is,
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of the Reflux Classifier 177
The channel aspect ratio is L/z = 1.0/0.006 = 166.19. Thus, the segregation
efficiency, η, is
1 1
η= = = 0.0731
0.33
1 + 0.133cos θ Ret L / z ( )
1 + 0.133cos 70° 4.790.33 166.19 ( ) ( )
The effective vessel area factor is
L
F = 1 + cos θ sin θ = 1 + 166.19cos ( 70° ) sin ( 70° ) = 54.41
z
The area occupied by the boxes relative to that of the lower cylindrical cross sec-
tion is f = 0.79. Thus, the open area available for the inclined channels is 0.79
× 0.80 × 3.14 m2 = 2.0 m2.
Therefore, the actual throughput factor = fαU/Ut = 0.79 × 0.80 × 3.98
= 2.53. Hence, this system should process at a solids feed rate 2.5 times that
of a TBS having the same lower cross-sectional area. Note that when finer
particles are targeted, this throughput advantage will increase markedly due
to the particle Reynolds number in Equation 3. For example, if the relevant
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
178 Beneficiation Technologies
The hindered settling that exists within the inclined channels is a complex
question, outside the scope of this chapter and the simple design approach
outlined here. It is assumed that the volume fraction of the suspended par-
ticles within the inclined channels is the same as the level in the overflow that
emerges. This assumption is reasonable for a system operating close to maxi-
mum capacity.
For systems incorporating particles of different density, the level of hin-
dered settling is best determined using the equation first proposed by Asif
(1997) given by
n
ρ − ρsusp
h= p (EQ 9)
ρp − ρ f
where n is 4.5. This exponent should arguably be smaller in value when these
larger particles are involved; however, in practice, a significant fraction of the
feed particles are finer than the size of the target particle, especially when slimes
are present. The effect of slimes is best assessed by introducing a modified fluid
density and viscosity into the calculation of the particle Reynolds number
and hence terminal velocity. Additional hindered settling is then applied via
Equation 9 based on the suspension density produced by the slimes and by the
nonslimes particles.
The hindered-settling factor will be lower for the mineral matter than for
the coal. This fact explains the autogenous dense-medium effect responsible for
preventing low-density coal from reporting to the underflow and also underpins
the inversion phenomenon of fluidized beds (Di Felice 1995). Here ρs is the
density of the target particle, 1,900 kg/m3, ρsusp is the suspension density of the
suspended solids in the inclined channels, and ρf is the density of water. In these
calculations the volume fraction of solids suspended in the inclined channels
is specified as φ = 0.2. Although a higher concentration is possible, this level
is regarded as boarding on the maximum. Assuming the solids transported up
through the inclined channels have an average solids density ρs = 1,400 kg/m3 for
a coal separation, then the suspension density is
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of the Reflux Classifier 179
Therefore, the hindered-settling factor applied to the inclined channels for the
target particle is
4.5
1900 − 1080
h= = 0.66
1900 − 1000
Thus, the hydraulic velocity applicable to the target particle, corrected for hin-
dered settling, is
This is the highest velocity permitted up through the vertical section of the
reflux classifier, just ahead of the inclined zone. Allowing for the vessel geom-
etry, the volumetric overflow rate can now be determined. The reference vessel
area used here is A = 3.14 m2. Thus the volumetric overflow rate is
It is assumed that the volume fraction of the solids in the overflow is equal
to that of the suspended particles within the inclined channels, φ = 0.2. This
is clearly not true for a system that is underloaded; however, it is a reasonable
assumption for a system operating at a throughput close to its maximum. Thus,
• The solids concentration of the overflow is C = φρs = 0.2 × 1,400 =
280 kg/m3
• The solids rate to the overflow is MO = QC = 363 × 280/1,000 =
101.5 t/h
• The water concentration of the overflow is C = (1 – φ)ρs = 0.8 × 1,000
= 800 kg/m3
• The water rate to the overflow is WO = 363 × 800/1,000 = 290.2 t/h
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
180 Beneficiation Technologies
The basic principal is to ensure that the bed is “just” suspended, and that
the rate used gives good stability. Insufficient fluidization will lead to a failure
in the measurement of the suspension density and hence a failure in the process
control, as well as the trapping of lower-density particles in the bed, and in turn
poorer Ep values. It is crucial to ensure that the bed is not excessively fluidized,
as this lowers the suspension density and induces unwanted mixing near the
underflow exit, while preventing fine dense particles from reaching the under-
flow. Clearly, by mass balance, the water rate with the feed solids is
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of the Reflux Classifier 181
Mo = 101.6
Wo = 290.2
MF = 119.5
WF = 229.7
RC
Fluidization Underflow
WD = 72.0 MU = 17.9
Wu = 11.5
Figure 11 Overall material balance of solids in t/h (M) and water (W) in the feed (F),
underflow (U), and overflow (O), with fluidization, D
about 30% solids for coal and mineral matter feeds, given this level allows the
system to self-adapt to variable solids loadings, assuming the water rate remains
reasonably steady.
The overall material balance is shown in Figure 11 in terms of the mass of
solids and water in each of the flow streams. Notably, if the target particle size
is 0.25 mm, a more conservative design would be based on a smaller size in the
calculation, perhaps 0.125 mm, if very high performance is sought. This will
drive the solids throughput down by a factor of about 2.0 to 60 t/h. Further,
in the presence of slimes, involving particles significantly finer than 0.25 mm,
hindered settling could be more significant. The level of hindered settling can
vary appreciably depending on the extent to which the clays are space filling.
Thus, care is required in estimating the effects of slimes.
It is appropriate to end with some comment on the application of the lami-
nar shear mechanism to the design calculations. This mechanism is primarily
concerned with providing a basis for the conveying of particles to the overflow.
Application of the theory of Galvin and Liu (2011) to a channel with z = 6.0
mm, and particles covering the range d = 0.25 to 2.0 mm, and density 1,900
kg/m3, produces a range of channel velocities, U '. An example calculation can
be found in the paper. Note, however, that the particle terminal velocity in
both the normal and tangential directions should be lowered by the relevant
hindered-settling factor, 0.66. This approach reduces the required inertial lift
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
182 Beneficiation Technologies
force and local fluid velocity at the inclined surface. The value of U ' is then
reduced by the area correction factor of 0.79 × 0.80 to allow for the plate
thickness, and launders. A further factor of sin70° is applied to convert the
channel velocity to the vertical direction, based on continuity considerations.
Equation 6, when modified by these factors, gives the approximation U = [Ut z/
(3d)] × [0.66 × 0.79 × 0.80 × sin(70°)] = [0.0192 × 6/(3 × 0.25)] × 0.39 =
0.06 m/s. The value based on the exact velocity profile is 0.052 m/s. The more
complete theory, incorporating inertial lift, results in a velocity of 0.036 m/s for
0.5-mm-diameter particles, 0.032 m/s for 1.0-mm particles, 0.038 m/s for 1.5-
mm particles, and 0.045 m/s for 2.0-mm particles. This trend has similarities
to the curve shown in Figure 9, with an average value of 0.041 clearly higher
than the value of 0.0321 m/s obtained using Equations 3, 4, and 5. The aver-
age value obtained using a lower target density of 1,700 kg/m3 is 0.031 m/s,
which is lower than 0.0321 m/s, so there is reasonable consistency between
the methods. The laminar shear mechanism tends to favor the transport of the
lower-density particles, regardless of their size.
Moreover, the suspension velocity alone does not govern the actual sepa-
ration. Rather, it is a combination of the inclined channel velocity and level
of hindered settling that matters. These two values are controlled by the set
point density applied to the lower bed. When the set point density is exceeded,
underflow discharges, reducing the inclined channel velocity. Hence, in prac-
tice, the upper inclined channel zones self-adapt, building the volume fraction
of the solids in the inclined channels to a level sufficient for the transport of
the particles to the overflow. The lowest density particles will always tend to
transport ahead of the higher-density particles. An excessive velocity will lead
to a lower volume fraction of solids in the inclined channels, whereas a lower
velocity will lead to a higher volume fraction. Conversely, a relatively low solids
throughput will lead to the building of the volume fraction of solids within the
inclined channels, though the volume fraction in the actual overflow emerging
from the unit will be low. In general, the holdup, or volume fraction of solids,
required to convey the particles to the overflow will develop automatically, with
the lower-density particles always in the “front of the queue.”
CONCLUSIONS
This chapter outlined the development of the reflux classifier, from inception,
through the initial pilot-scale phase, to the development of the first prototype.
The importance of the ongoing research and subsequent advance achieved
when closely spaced inclined channels were used, were also described. This
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of the Reflux Classifier 183
ACKNOWLEDGMENTS
The author acknowledges the financial support of the Australian Research
Council and the Australian Coal Association Research Program, the significant
contributions of Ludowici Australia, and the ongoing support of the Bloom-
field Group in hosting plant trials.
REFERENCES
Acrivos, A., and Herbolzheimer, E. 1979. Enhanced sedimentation in settling tanks
with inclined walls. J. Fluid Mech. 92(3):435–457.
Asif, M. 1997. Modeling of multi-solid liquid fluidized beds. Chem. Eng. Technol.
20:485–490.
Bird, B., Stewart, W., and Lightfoot, E. 1976. Transport Phenomena. New York: John
Wiley and Sons. p. 62.
Boycott, A.E. 1920. Sedimentation of blood corpuscles. Nature 104:532.
Carty, J.J. 1957. Resistance coefficients for spheres on a plane boundary. B.S. the-
sis, Department of Civil and Sanitary Engineering, Massachusetts Institute of
Technology.
Davis, R.H., Zhang, X., and Agarwala, J.P. 1989. Particle classification for dilute sus-
pensions using an inclined settler. Ind. Eng. Chem. Res. 28:785–793.
Di Felice, R. 1995. Hydrodynamics of liquid fluidization. Chem. Eng. Sci. 50:1213–1245.
Doroodchi, E., Fletcher, D.F., and Galvin, K.P. 2004. Influence of inclined plates on
the expansion behaviour of particulate suspensions in a liquid fluidised bed. Chem.
Eng. Sci. 59:3559–3567.
Doroodchi, E., Zhou, J., Fletcher, D., and Galvin, K.P. 2006. Particle size classification
in a fluidised bed containing parallel inclined plates. Miner. Eng. 19:162–171.
Galvin, K.P. 2004. A reflux classifier. University of Newcastle Research Associates Lim-
ited. U.S. Patent 6814241.
Galvin, K.P., and Liu, H. 2011. Role of inertial lift in elutriating particles according to
their density. Chem. Eng. Sci. 66:3686–3691.
Galvin, K.P., and Munro, M. 2008. Overflow launder. University of Newcastle Research
Associates Limited. U.S. Patent 7334689.
Galvin, K.P., and Nguyentranlam, G. 2002. Influence of parallel inclined plates in a
liquid fluidized bed system. Chem. Eng. Sci. 57:1231–1234.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
184 Beneficiation Technologies
Galvin, K.P., Pratten, S.J., and Nicol, S.K. 1999. Dense medium separation using a
teetered bed separator, Miner. Eng. 12(9):1059–1081.
Galvin, K.P., Zhao, Y., and Davis, R.H. 2001. Time-averaged hydrodynamic roughness
of a noncolloidal sphere in low Reynolds number motion down an inclined plane.
Phys. Fluids 13(11):3108–3119.
Galvin, K.P., Doroodchi, E., Callen, A.M., Lambert, N., and Pratten, S.J. 2002. Pilot
plant trial of the reflux classifier. Miner. Eng. 15:19–25.
Galvin, K.P., Callen, A., Zhou, J., and Doroodchi, E. 2004. Gravity separation using a
full-scale reflux classifier. In Proceedings of the Tenth Australian Coal Preparation
Conference. Edited by W.B. Membrey. Paper H21.
Galvin, K.P., Callen, A., Zhou, J., and Doroodchi, E. 2005. Performance of the reflux
classifier for gravity separation at full scale. Miner. Eng. 18:19–24.
Galvin, K.P., Walton, K., and Zhou, J. 2009. How to elutriate particles according to
their density. Chem. Eng. Sci. 64:2003–2010.
Galvin, K.P., Walton, K., Zhou, J. 2010. Application of closely spaced inclined channels
in gravity separation of fine particles. Miner. Eng. 23:326–338.
King, M.R., and Leighton, D.T. 1997. Measurement of the inertial lift on a moving
sphere in contact with a plane wall in shear flow. Phys. Fluids 9(5):1248–1255.
Laskovski, D., Duncan, P., Stevenson, P., Zhou, J., and Galvin, K.P. 2006. Segrega-
tion of hydraulically suspended particles in inclined channels. Chem. Eng. Sci.
61:7269–7278.
Leighton, D., and Acrivos, A. 1986. Viscous resuspension. Chem. Eng. Sci.
41(6):1377–1384.
MacTaggart, R.S., Him-Sum Law, D., Masliyah, J.H., and Nandakumar, K. 1988. Grav-
ity separations of concentrated bidisperse suspensions in inclined plate settlers.
Int. J. Multiphase Flow 14:519–532.
Nakamura, H., and Kuroda, K. 1937. La cause de l’acceleration de la vitesse de sedi-
mentation des suspensions dans les recipients inclines. Keijo J. Med. 8:256–296.
Ponder, P. 1925. On sedimentation and Rouleaux formation. Q. J. Exp. Physiol.
15:235–252.
Rampall, I., and Leighton, D.T. Jr. 1994. Influence of shear-induced migration on tur-
bulent resuspension, Int. J. Multiphase Flow 20(3):631–650.
Richardson, J.F., and Zaki, W.N. 1954. Sedimentation and fluidization: Part I. Trans.
Inst. Chem. Eng. 32:35–53.
Smart, J.R., Beimfohr, S., and Leighton, D.T. 1993. Measurement of the translational
and rotational velocities of a noncolloidal sphere rolling down a smooth inclined
plane at low Reynolds number. Phys. Fluids A 5:13.
Thompson, P.D., and Galvin, K.P. 1997. An empirical description for the classification
in an inclined counter-flow settler. Miner. Eng. 10(1):97–109.
Vance, W.H., and Moulton, R.W. 1965. A study of slip ratios for the flow of steam-
water mixtures at high void fraction. AIChE J. 11(6):1114–1124.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of the Reflux Classifier 185
Wills, B.A., and Napier-Munn, T.J. 2006. Mineral Processing Technology: An Introduc-
tion to Practical Aspects of Ore Treatment and Mineral Recovery. 7th ed. London:
Butterworth-Heinemann.
Yao, K.M. 1973. Design of high-rate settlers. J. Environ. Eng. Div. 99(5):621–637.
Zhang, X., and Davis, R.H. 1990. Particle classification using inclined settlers in series
and with underflow recycle. Ind. Eng. Chem. Res. 29:1894–1900.
Zhou, J., Walton, K., Laskovski, D., Duncan, P., and Galvin, K.P. 2006. Enhanced
separation of mineral sands using the reflux classifier. Miner. Eng. 19:1573–1579.
Zigrang, D.J., and Sylvester, N.D. 1981. An explicit equation for particle settling veloci-
ties in solid-liquid systems. AIChE J 27(6):1043–1044.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Column and Nonconventional
Flotation for Coal Recovery:
Circuitry, Methods, and
Considerations
J.N. Kohmuench, E.S. Yan, and L. Christodoulou
ABSTRACT
Column flotation continues to be an efficient and economical means of recovering
fine coal from wash plant effluent streams. The use of deep froths and the applica-
tion of wash water ensure a high-quality float product. As well, the use of robust
sparging systems ensures high combustible recoveries through the generation of copi-
ous amounts of fine bubbles. With more than a dozen years of supplying flotation
systems for the coal industry, Eriez Manufacturing Company continues to gain
insight into the proper application of these devices and how the flotation circuit
interacts with the remainder of the plant circuitry. This understanding is essential
as a handful of specific parameters can have a large impact on several aspects of
the design of the flotation circuit including, but not limited to, unit equipment
selection and sizing, circuit arrangement, layout, up- and downstream ancillary
requirements, ease of operation, and ultimately capital cost. Incorporating these
findings and taking advantage of new technology and approaches can help end
users achieve the desired rate of return for a flotation project.
INTRODUCTION
Flotation columns derive their name from the geometric shape of the vessel.
Unlike conventional mechanically agitated flotation machines, which tend to
use relatively shallow rectangular tanks, column cells used in the coal industry
are tall vessels with heights typically ranging from 7 m to 16 m depending on
the application. Unlike conventional flotation machines, columns do not use
mechanical agitation. The absence of intense agitation promotes higher degrees
of selectivity and aids in the recovery of coarse particles. In general, feed slurry
187
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
188 Beneficiation Technologies
enters the column at one or more feed points located in the upper third of the
column body and descends against a rising swarm of fine bubbles generated by
the air sparging system. Particles that collide with and attach to the bubbles rise
to the top of the column, eventually reaching the interface between the pulp
(collection zone) and the froth (cleaning zone).
For an equivalent volumetric capacity, the cross-sectional surface area of
the column cell is much smaller than a conventional cell. This reduced area
is beneficial for promoting froth stability and allowing deep froth beds to be
formed. This is an important aspect of column flotation, as a deep froth bed
facilitates froth washing to remove unwanted impurities from the float product.
Wash water, added at the top of the column, percolates through the froth zone
displacing dirty process water and nonselectively entrained particles trapped
between the bubbles. In addition, froth wash water serves to stabilize and add
mobility to the froth. Sufficient water must be added to ensure that all of the
feedwater that normally reports to the froth product has been replaced with
fresh or clarified water. Less than 1% of the feed pulp and associated clays will
report to the froth in a well-operated column (Luttrell et al. 1999).
In contrast, conventional mechanical cells do not operate with deep froths.
Therefore, these devices allow some portion of the ultrafine mineral slimes to be
recovered with the water that reports to the froth, consequently reducing prod-
uct quality. This particle entrainment is the nonselective hydraulic conveyance
of gangue into the product launder. In fact, fine particles (<0.045 mm) have a
tendency to report to the froth concentrate in direct proportion to the amount
of product water recovered. As such, the flotation operator is often forced to
make the decision to either “pull hard” on the cells to maintain yield or run the
cells less aggressively to maintain grade.
The primary advantage of utilizing wash water is the ability to provide a
superior separation performance compared to conventional flotation processes.
This capability is illustrated by the test data summarized in Figure 1, which
compares column flotation technology with an existing bank of conventional
cells. As shown, the separation data for the column cells utilizing wash water
are far superior to those obtained from the conventional flotation bank. In fact,
the data for the column cells tend to fall just below the separation curve pre-
dicted by release analysis (Dell et al. 1972). A release analysis is an indication
of the ultimate flotation performance and is often regarded as “washability” for
flotation. This figure suggests that columns provide a level of performance that
would be difficult to achieve even after multiple stages of cleaning by conven-
tional machines.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Column and Nonconventional Flotation for Coal Recovery 189
100
Release
Analysis
80
60
Recovery, %
40
20
Full-Scale Conventional
Laboratory Column
Full-Scale Column
0
0 5 10 15 20
Product Ash Content, %
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
190 Beneficiation Technologies
0.15 mm (Norm.)
Column Feed
Filtrate
Adjustable
Wash-Water
Ring
To Spirals
or Clean Clean Coal
Coal De-aeration
Dewatering Tank
Circuit
Feed
of clay that is recovered into the flotation launder. As such, there is significant
economic incentive to minimize entrainment and improve plant performance.
T Y P I C A L F I N E C O A L C O L U M N F L O TAT I O N C I R C U I T S
There are two typical circuits utilized for fine coal flotation. These include
the traditional minus-0.150-mm “by-zero” circuit and the 0.150 × 0.045 mm
“deslime” circuit. As would be expected, the by-zero approach typically incor-
porates the use of large-diameter cyclones that treat minus-1-mm feed as seen
in Figure 2. Typically, these cyclones are configured to achieve a cut point of
approximately 0.150 mm with the finer fraction reporting directly to flotation.
In some markets, the cut point can be as coarse as 0.5 mm, though it is recom-
mended to not exceed 0.250 mm.
The feed to flotation in a by-zero circuit is relatively dilute and typically
ranges between 2.5% and 6.0% solids, by weight. As a result, volume flows in
these circuits can be quite large. Consequently, significant cell volume must be
available to achieve the desired retention time to ensure adequate combustible
recovery. To attain the needed retention time, a number of large-diameter cells
may be required, which can range in height from 8.5 m to 16.0 m. The ultimate
height is typically based on kinetic data generated from laboratory testing, but
it is generally constrained by practical limits as dictated by both economic and
site-specific engineering requirements. As a result, the height-to-diameter ratio
for an industrial cell ranges from 3:1 to as low as 2:1.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Column and Nonconventional Flotation for Coal Recovery 191
100
95
90
85
80
Recovery, %
75
70
65
60 Plug Flow
Two Cells in Series (3.2 Mixers)
Two Cells in Parallel (1.6 Mixers)
55
50
0 1 2 3 4 5 6 7 8 9 10
kτ
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
192 Beneficiation Technologies
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Column and Nonconventional Flotation for Coal Recovery 193
Reject
Recycled
Screen-bowl
Effluent
Clean Coal
To Spirals 0.150 × 0.045 mm De-aeration
or Clean (Nominal) Tank
Coal
Screen Bowl
Dewatering
Circuit
Feed
coarser float product and a significant reduction in total frother addition. The
coarser froth product tends to be less stable and will readily collapse. The lower
frother addition rate is a result of the volume split achieved in the smaller 150-
mm cyclones, which substantially reduces the total volume of slurry that must
be processed in flotation. Given that frother is added based on total flotation
volume flow, this allows high concentrations of surfactant to be realized in the
flotation circuit without causing upsets elsewhere in the plant. The obvious sec-
ondary benefit is the reduction in size and cost of the flotation and downstream
dewatering equipment. Overall, Bethell and Luttrell (2005) indicate that a
deslime circuit may cost less than half of an equivalent by-zero circuit.
CELL TECHNOLOGY
The ability to consistently maintain high combustible recoveries while pro-
ducing low ash flotation product is the result of operating with efficient and
robust sparging systems. Furthermore, the high grade float product is a result of
operating with a deep froth while incorporating the generous addition of wash
water. In order to run with deep froths, the sparging systems must be designed
to provide the maximum rate of bubble surface area through the column. As
such, the air sparging system is perhaps the most important component in a
column flotation cell. While details related to the specific design features of the
various sparging technologies have been presented in the literature (McKay et
al. 1988; Davis et al. 1995; Finch 1995), the most prevalent column sparging
systems are presented here.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
194 Beneficiation Technologies
SlamJet Technology
The first practical bubble generation system was developed by the U.S. Bureau
of Mines. This design essentially consisted of a series of small orifices (2–3 mm)
that were placed along the length of a pipe which was then inserted through
the column sidewall. This approach worked satisfactorily but was plagued with
maintenance problems with regard to plugging of the orifices. Canadian Pro-
cess Technologies (CPT) solved the plugging problems with the introduction
of their SparJet aeration system. This system uses a series of removable air lances
that include a single orifice located at the end of each sparger. High-velocity
air is injected into the column cell to create and disperse fine bubbles. This
technology was further refined when CPT, now the Eriez Flotation Division
(EFD), updated the SparJet spargers by incorporating a self-closing mechanism
that eliminates any backflow of slurry into the aeration system in the event of
an air supply failure. This updated version has been designated the SlamJet and
is presented in Figure 5.
SlamJet technology has best been applied to the coarser 0.150 × 0.045 mm
deslime flotation circuits where the coarser feed does not require intense colli-
sion energy to achieve particle collection. In addition, the lower volumetric feed
rate provided by this circuit ensures recovery by providing significant retention
time that can easily exceed 10–12 minutes in a typical deslime column that is
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Column and Nonconventional Flotation for Coal Recovery 195
fed between 10% and 14% solids, by weight. The bubble size distribution pro-
vided by this sparging technique is well suited for this size class and provides
efficient bubble generation powered only by compressed air.
Cavitation-Tube Technology
One of the most popular sparging technologies used in the coal industry is
generally considered the dynamic sparging system. This technology was con-
ceived after several years of fundamental research which showed that the rate
of flotation could be enhanced through the use of smaller air bubbles. The small
(<0.8 mm) bubbles are generated by circulating slurry from the lower section
of the column through a set of parallel, in-line spargers into which compressed
air is injected (Figure 6). The spargers are typically mounted outside the col-
umn to simplify inspection and replacement. The dynamic sparging system has
been best applied on by-zero fine coal circuits. The primary reason is that the
dynamic sparging system has the ability to efficiently generate large amounts
of very small bubbles for a given airflow. As such, the probability of bubble–
particle collisions is greatly improved, which can increase capacity. This is an
added benefit when treating coal in a traditional by-zero circuit, which has a
high-volume flow, reduced residence time, and a finer feed size distribution.
More recently, CPT developed and implemented the Cavitation-Tube
sparger which can meet the rigorous performance demands of industrial instal-
lations. This sparger consists of an hourglass-shaped tube (Figure 6) constructed
of wear-resistant material. As part of a dynamic sparging system, a mixture of
flotation pulp and compressed gas is circulated at very high velocity through
the device using a centrifugal pump. The high velocity and throat geometry
creates cavitation in the slurry that is beneficial for the flotation of ultrafine
particles (Zhou et al. 1994, 2009). The gas nuclei that are created during cavita-
tion help to activate particles by promoting the attachment of larger bubbles.
In other words, the tiny bubbles act as a secondary collector, which can reduce
flotation collector dosage and enhance the probability of particle attachment.
Details and further descriptions of the advantages of cavitation in flotation
have been described elsewhere (Santos et al. 2012; Honaker et al. 2011).
StackCell
Over the last several years, EFD has developed a new flotation device called
the StackCell that offers high flotation capacity in a small footprint. This
approach provides column-like performance while offering the flexibiilty to
arrange by-zero flotation circuits in such a manner to optimize both metallur-
gical and economical efficiency. Although a detailed theory of operation and
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
196 Beneficiation Technologies
Adjustable
Wash-Water
Distributor
Froth Launder
Feed Inlet
Froth Product
Slurry
Manifold
Slurry
Recirculation
Pump
Cav-Tube
Air
Manifold Tailing
Discharge
Valve
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Column and Nonconventional Flotation for Coal Recovery 197
Aeration/
Contacting
Chamber Wash-Water
Manifold
Froth
Launder
Feed Inlet
Tails
Outlet
Air Manifold
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
198 Beneficiation Technologies
100
80
Recovery, %
60
Avg = 74.4% Avg = 83.7%
40
20
0
100
80
Reject Ash, %
60
20
0
0 50 100 150 200 250 300
Sample Number
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Column and Nonconventional Flotation for Coal Recovery 199
Table 1 Various flotation options for 70-tph flotation by-zero circuit with 80% float
yield
StackCell Column Combination
Circuit Variable SC-3050 Cav-Tube SC-3050 Cav-Tube
Cell diameter, m 3.00 4.25 3.00 4.25
Cell, no. 4 4 2 2
Circulation pump — P — P
Blower P — P —
Compressor — P — P
Total connected power, hp 840 1,600 1,110
Live load, kt 404 1,140 671
Estimated equipment cost, US$ 1.5M 1.5M 1.2M
P I T FA L L S A N D B E S T P R A C T I C E S
Over the past decade, Eriez has continued its research and development efforts
with regard to circuit design and column sizing. Important design parameters
include product carrying capacity, retention time, aeration rate, wash-water
rate, and chemical dosage. Many of these parameters are determined from
either feed characterization studies and lab- or pilot-scale test data. However,
special consideration should be given to the following parameters to ensure that
the flotation circuit can be operated at maximum efficiency.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
200 Beneficiation Technologies
Carrying Capacity
Carrying Capacity
Feed Particle Size Ultrafines Content
Figure 9 Correlation between capacity, mean size, and ultrafines content for coal
flotation
Froth Washing
The use of wash water in a column or StackCell is important and necessary to
provide the optimum separation efficiency as indicated by release analysis. In
addition, a froth depth of 0.6–1.2 m is typically required to ensure good distri-
bution of the wash water and to prevent short-circuiting. As with any column,
the flow of wash water must exceed the volumetric flow of water reporting to
the clean coal product to prevent entrainment of the high-ash slimes. In most
cases, less than about 1% of the feedwater will report to the froth product if the
wash water is properly controlled. The amount of water carried by the froth can
be calculated from:
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Column and Nonconventional Flotation for Coal Recovery 201
20
16
14
12
Froth Ash, %
10
Typical
2 Operation
0
0 1 2 3 4
Number of Dilution Washes
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
202 Beneficiation Technologies
surfaces. High water rates may also have a detrimental impact on product grade
by increasing axial froth mixing, thereby reducing the wash-water effectiveness
(Yianatos et al. 1988).
The design of the wash-water distributor can also significantly affect col-
umn performance. In some cases, the distribution piping is intentionally sub-
merged below the cell lip so that a drained froth can form above the distributor.
This arrangement allows for exacting control of the water split between the
clean coal and the refuse streams. Ultimately, changes to the vertical position of
the distributor can be used to control the product grade. In some cases, multi-
level concentric distribution rings can be used to overcome problems associated
with poor froth mobility. The inner rings are typically located above the outer
rings to reduce drainage and improve the fluidity of the froth in the center of
the column. More recently, one or more internal launders have been employed
in lieu of tiered wash-water rings. Unfortunately, it can be difficult to identify
and clean out plugged distributor piping, which can severely impact cell per-
formance. As a result, static wash pans have been employed that are located just
above the top of the froth. This arrangement does not provide froth mobility
control, but does allow for easy cleaning and reduced maintenance require-
ments given that they can be easily cleared using a plant water hose.
Aeration
The primary advantage of an efficient and commercially available sparging
system is the ability to generate large amounts of very small bubbles. This capa-
bility is commonly reported in terms of the superficial bubble surface area rate
(Sb), which is defined as the total bubble surface area per unit of time passing
through a given column cross-sectional area. This value can be calculated by
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Column and Nonconventional Flotation for Coal Recovery 203
Figure 11 Improper air instrumentation arrangement with no pressure regulator and
with air flowmeter incorrectly installed following the flow control valve
and bubble size will generally provide a gas holdup in the flotation pulp above
15%–18%. Values above 20% are not unusual for well-tuned circuits.
Caution should be used during the metering of gas flow rates. A properly
designed system should be equipped with a flowmeter that is calibrated to
read correctly at a specified operating pressure. The operating pressure should
be held constant by placing a pressure regulator ahead of the flowmeter and
followed by a control valve so that the meter will always operate at its design
pressure. If the flowmeter is placed after the control valve, then the operating
pressure is unknown and the true gas flow rate cannot be determined (see Fig-
ure 11). In addition, the air supply lines should be as large as reasonably possible
to minimize line losses. Supply lines with an excessive number of fittings, mul-
tiple changes in direction, and small pipe diameters can severely limit the total
volume and pressure of air that can be delivered to the flotation cell.
A great deal of confusion also exists regarding the specification of compres-
sors for column applications. Much of this confusion is related to improper
use of gas flow terminology (Sullair Corporation 1992). For example, column
manufacturers normally report gas flow rates as a “standard” volumetric flow
per time. This value is only valid at 1 atm of pressure and 20°C of dry air. The
“actual” flow rate specified by compressor manufacturers is typically reported
in terms of “inlet” conditions or “free air.” Although this amount of air enters
the compressor, it is not necessarily the amount of air delivered to the column
due to compressor seal leakage. As a result, the actual flow may be only 95% of
the inlet flow. Furthermore, corrections to the gas flow rate must be made to
account for differences in elevation (atmospheric pressure) and humidity. Air
temperature generally has little impact on the capacity of an oil-flooded screw
compressor, but it may affect the performance of an air-cooled compressor.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
204 Beneficiation Technologies
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Column and Nonconventional Flotation for Coal Recovery 205
this problem by selecting weaker frothers or reducing frother dosage have not
been successful and have generally led to lower circuit recoveries.
Several circuit modifications have been developed to deal with the froth
stability problem. For example, column launders need to be considerably
oversized with steep slopes to reduce backup. Horizontal froth travel distances
must be kept as short as possible and adequate vertical head must be provided
between downstream operations and column launders. In addition, piping and
chute work must be designed such that the air can escape as the froth travels
from the flotation circuit to the next unit operation. Figure 12 shows how
small changes in piping arrangements can result in better process performance.
Figure 12a shows a column whose performance suffered because of the inability
to move the froth product from the column launder although a large discharge
nozzle (1 × 1 m) had been provided. In this example, the froth built up in the
launder and overflowed when the operators increased air rates. To prevent this,
the air rates were lowered, which resulted in less-than-optimum recoveries.
It was determined that the downstream discharge piping was air-locking and
preventing the launders from properly draining. The piping was replaced with
larger chute work that allowed the froth to flow freely and the air to escape. As a
result, higher air rates were possible and recoveries were significantly improved.
Some installations have resorted to using defoaming agents or high-
pressure launder sprays to deal with froth stability. However, newer column
installations avoid this problem by including large deaeration tanks to permit
time for the froth to collapse (Figure 13a). Special provisions may also be
required to ensure that downstream dewatering units can accept the large froth
volumes. For example, standard screen-bowl centrifuges equipped with 10-cm
(4-in.) inlets may need to be retrofitted with 20-cm (8-in.) or larger inlets to
minimize flow restrictions.
The use of screen-bowl centrifuges provides low product moistures; how-
ever, there are typically fine coal losses as a large portion of the –0.045 mm float
product is lost as main effluent. This material is highly hydrophobic and will
typically accumulate on top of the thickener as a very stable froth layer, which
increases the probability that the process water quality will become contami-
nated (i.e., black water). This phenomenon is more prevalent in by-zero circuits,
especially when the screen-bowl screen effluent is recycled back through the
flotation circuit, either directly or through convoluted plant circuitry. Reintro-
ducing material that has already been floated to the flotation circuit can result
in a circulating load of very fine and highly floatable material. As a result, the
capacity of the flotation equipment can be significantly reduced, which results
in losses of valuable coal. Most installations will combat this by ensuring that
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
206 Beneficiation Technologies
A B
Figure 12 Difference in froth discharge before and after changes to downstream
piping
A B
Figure 13 (a) Column with deaeration tank and (b) thickener with floating boom and
sprays
the screen-bowl screen effluent is routed directly back to the screen bowl so
that it does not return to the flotation circuit. The accumulation of froth on the
thickener is also reduced by utilizing reverse weirs and taller center wells, as this
approach helps limit the amount of froth that can enter into the process water
supply. Froth that does form on top of the clarifier can be eliminated by employ-
ing a floating boom that is placed directly in the thickener (Figure 13b) and
used in conjunction with water sprays. The floating boom is constructed out of
inexpensive PVC (polyvinyl chloride) piping that can be attached to the rotat-
ing rakes. The boom floats on the water interface and drags any froth around to
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Column and Nonconventional Flotation for Coal Recovery 207
the walkway that extends over the thickener. Water sprays are mounted on the
walkway and positioned to maximize the destruction of the froth.
For very coarse feeds, the froth product may be sufficiently handleable
to be pumped to classifying cyclones. In such cases, the cyclone underflow
(oversize) product is typically passed directly to the dewatering circuit (usually
filters), while the cyclone overflow (undersize) product is passed to a clean coal
thickener. The solids from the thickener underflow are pumped to the dewater-
ing circuit. The clarified overflow from the thickener is sent back to the column
circuit as wash water. This arrangement maintains a high frother concentration
in the column circuit and reduces the amount of fresh frother that is required,
thereby minimizing the impact of surfactant buildup on other plant circuits.
This approach can also be mimicked using vacuum belt and disk filters by rout-
ing the relatively clean filtrate back to the flotation circuit to be used as wash
water.
SUMMARY
As high-grade feed stocks continue to decline, the role of flotation will likely
increase in future plant designs. As such, it is important to understand how
both column and other nonconventional flotation can be applied to maximize
profitability. This includes ensuring that the circuit design and equipment selec-
tion compliment the product requirements and how they ultimately relate to
the coal sales contract. Experience has shown that the final market destination
(i.e., metallurgical vs. thermal) and contract specifications (i.e., moisture, tons,
and inerts) will be the primary factors when making these process decisions.
Although the installation of a column or StackCell circuit can provide a
beneficial financial return, the design and scale-up of this technology is chal-
lenging. Outside issues such as froth handling and dewatering techniques
must be fully considered to ensure a successful outcome. Careful engineering,
extensive testing, and attention to ancillary requirements will ensure that the
flotation performance agrees with that predicted by standard release analysis.
To ensure the success of the entire installation, the overall design of the flo-
tation circuitry must also be taken in account. This includes the determination
of the circuit type (i.e., by-zero or deslime) and whether or not the cells should
be placed in parallel or in series. By itself, the selection of a flotation column
or a StackCell requires the understanding of application-specific details, such
as coal quality and expected product mass yield. Through this understand-
ing, the cells can be properly engineered with regard to sparging technology,
product carrying capacity, retention time, wash-water rates, and aeration rates.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
208 Beneficiation Technologies
Adherence to the design criteria as outlined in this manuscript will provide the
basis for a successful flotation circuit.
REFERENCES
Bethell, P.J., and Luttrell, G.H. 2005. Effects of ultrafine desliming on coal flotation
circuits. In Proceedings, Centenary of Flotation Symposium, Brisbane, Australia. pp.
719–728.
Davis, V.L., Bethell P.J., Stanley F.L., and Luttrell G.H. 1995. Plant practices in fine coal
column flotation. High Efficiency Coal Preparation: An International Symposium.
Edited by S.K. Kawatra. Littleton, CO: SME. pp. 237–246.
Davis, V., Stanley, F., Kiser, M., Bratton, R., Luttrell, G.H., Yan, E.S., Kohmuench,
J.N., and Christodoulou, L. 2011. Industrial evaluation of the StackCell flotation
technology. Coal Prep. Soc. Am. (CPSA) J. 10(3):22–26.
Dell, C.C., Bunyard, M.J., Rickelton, W.A., and Young, P.A. 1972. Release analysis: A
comparison of techniques. Trans. IMM (Sect. C), 81(787):89–96.
Finch, J.A. 1995. Column flotation: A selected review—Part IV: Novel flotation
devices. Miner. Eng. 8(6):587–602.
Honaker, R.Q., Saracoglu, M., Kohmuench, J.N., and Mankosa, M.J. 2011. Cavitation
pretreatment of a flotation feedstock for enhanced coal recovery. 28th Interna-
tional Coal Preparation Conference and Exhibit, Lexington, KY, May 2–5.
Kohmuench, J.N., Davy, M.S., Ingram, W.S., Brake, I.R., and Luttrell, G.H. 2004.
Benefits of column flotation using the Eriez Microcel. In Tenth Australian Coal
Preparation Conference, Proceedings, Polkolbin, NSW, Australia, October 17–21.
pp. 272–284.
Kohmuench, J.N., Mankosa, M.J., and Yan, E.S. 2008. An alternative for fine coal flota-
tion. Coal Prep. Soc. Am. (CPSA) J. 7(1):29–39.
Levenspiel, O. 1972. Chemical Reaction Engineering. New York: Wiley.
Luttrell, G.H., Kohmuench, J.N., Stanley, F.L., and Davis, V.L. 1999. Technical and
economic considerations in the design of column flotation circuits for the coal
industry. In Proceedings, SME Annual Meeting and Exhibit, Symposium Honoring
M.C. Fuerstenau. SME Preprint 99-166. Denver, CO: SME.
Luttrell, G.H., Keles, S., and Honaker, R.Q. 2009. Implications of constant incremen-
tal quality on the design of fine coal dewatering circuitry. SME Preprint 09-093.
Denver, CO: SME.
McKay, J.D., Foot, D.G., and Shirts, M.D. 1988. Column flotation and bubble genera-
tion studies at the Bureau of Mines. In Column Flotation’ 88. Littleton, CO: SME.
pp. 173–186.
Santos, N.C., Duarte, F., Fan, M., Honaker, R.Q., Baldessin, A., and Pinheiro, C.P.
2012. Feed air jet—A cavitation sparger system for enhanced flotation recovery.
In Proceedings, 13th ABM Iron Ore Symposium, Rio de Janeiro, October 14–18.
Sastri, S.R.S. 1996. Technical note: Carrying capacity in flotation columns. Miner. Eng.
9(4):465–468.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Column and Nonconventional Flotation for Coal Recovery 209
Stanley, F., King, P., Horton, S., Kennedy, D., McGough, K., and Luttrell, G. 2006.
Improvements in flotation column recovery using cell-to-cell circuitry. In 23rd
International Coal Preparation Exhibition and Conference, Proceedings, Lexington,
KY, May 1–4.
Sullair Corporation. 1992. Sales bulletin. Industrial compressors—Compressor capac-
ity. In Bulletin E-78. April.
Yianatos, J.B., Finch, J.A., and LaPlante, A.R. 1988. Selectivity in column flotation
froths. Int. J. Miner. Process. 23:279–292.
Zhou, Z.A., Xu, Z., and Finch, J.A. 1994. On the role of cavitation in particle collection
during flotation—A critical review. Miner. Eng. 7(9):1073–1084.
Zhou, Z.A., Xu, Z., Finch, J.A., Masliyah, J.H., and Chow, R.S. 2009. On the role of
cavitation in particle collection during flotation—A critical review II. Miner. Eng.
22:419–433.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Mechanical Cells for Fine Coal
Flotation—An International
Perspective
Jeff Euston, Keith Wilkes, and Asa Weber
ABSTRACT
Froth flotation is currently the only practical and economic method for treatment of
the finest coal fractions. Flotation is a complex process calling on a range of chemi-
cal, hydrodynamic, and engineering disciplines. Fine coal usually comprises less
than 20% of the plant feed and often contains the highest ash. Dewatering is best
achieved by vacuum filtration, and even then, the final product may have relatively
high moisture content with the associated contractual and transport issues.
Thirty years ago, mechanical flotation circuits required large numbers of
small individual cells with high energy and operator demands. The introduction
of pneumatic and column flotation devices in the 1980s offered an apparently
simple solution to increasingly large banks of mechanical flotation cells. These new
technologies promised simplicity with reduced operator involvement in what was
essentially a single device. This simplicity has resulted in a compromise in terms of
combustibles recovery.
Mechanical flotation cells continue to play a major role in coal flotation around
the world, and there is evidence of a resurgence of interest in modern mechanical
flotation technologies. Increasing coal prices and concerns for the environment have
generated a new look at the opportunities for high recovery from the fines circuit.
This chapter summarizes the rationale behind the use of modern mechanical flota-
tion cells and describes their role in the modern coal preparation plant.
INTRODUCTION
Froth flotation has its origins at the beginning of the 20th century at Broken
Hill in Australia. In metalliferous applications, lower ore grades and larger
plants resulted in the need for increasingly fine grinding. Comminution
211
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
212 Beneficiation Technologies
followed by froth flotation quickly became the only practical process to recover
metalliferous ores.
At the time, coal preparation consisted largely of dry beneficiation tech-
niques, such as spiral pickers, techniques that had themselves taken over from
manual sorting from the middle of the 19th century onward. Froth flotation
was first applied to coal in laboratory testing in 1915 with the first operating
plant in England in 1920 and the first U.S. plant in 1930.
Legislation, particularly in the United Kingdom, to reduce the health
effects of airborne dust in coal mines and dry beneficiation plants resulted in
the use of copious volumes of water for dust suppression in underground mines.
This resulted in a wet feed to the preparation plant, reducing the efficiency of
the traditional dry techniques.
Wet processing techniques were developed, and these led to the generation
of a large volume of slurry containing small but significant coal values in addi-
tion to the shale and clay waste. This waste stream was not only a loss of valuable
fine coal, but it generated a large volume of what was essentially dirty water,
often referred to as blackwater. The excellent volume on flotation, published by
the Australasian Institute of Mining and Metallurgy (Lynch 2010), provides an
excellent review of flotation development and includes an enlightening chapter
on coal flotation.
The fine coal fraction typically comprises less than 20% of the coal plant
feed and represents the most difficult to treat stream from both a cost benefit
and processing approach, often with low coal values and high clay content.
Some of the negatives associated with coal flotation are listed here:
• The clays often associated with coal report with the fines, resulting in
high ash values for the ultrafine fraction.
• There is potential to add moisture to the final product, which can in
some operations lead to freezing and an increase in the calorific value
of the final product.
• Certain coal types containing oxidized coal and high clay content do
not respond well to froth flotation.
• Coal specifications may place a limitation on fines content.
The balance of these factors coupled with the historically low value of coal
has resulted in mixed attitudes toward coal flotation. Not only the choice of
technology, but whether or not coal flotation is worth the effort is by no means
clear cut. Increasing coal prices with the mining boom, which began in 2007,
generated a renewed interest in improving coal recovery across the full size
range. Even though the boom has not continued, awareness has been created of
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Mechanical Cells for Fine Coal Flotation 213
the coal values being lost to tailings. By the 1980s, coal producers who decided
to install froth flotation found themselves with large numbers of small cells
with individual mechanisms and manual control requirements and relatively
high energy requirements per metric ton of coal treated.
In 1996, an Australian ACARP report by Sanders (2006), summarized
by Heiser (1996), described mechanical flotation as suffering from a range of
issues:
• Highly varied design, with unclear advantages for each
• Poor air control, as observed from periodic extreme cell disturbance
• Unpredictable scale-up performance
• Continual operator/metallurgical attention
• Adversely affected as wear progresses
• No universally applicable control strategy
• Difficult to set up for optimization of recovery and grade
• Energy and reagent costs
At that time, innovative designs such as the Jameson cell and conventional
column devices promised to replace these large banks of mechanical cells with
simple devices with no moving parts, and having the potential for lower prod-
uct ash values and better theoretical grade/yield performance with a minimum
of operator intervention. In Australia, between 1982 and 2012, no complete
mechanical flotation circuits for coal were installed, although several expan-
sions did use mechanical cells. The convenience of these new processes was
seen as an acceptable compromise to produce a low-ash product. Elsewhere
around the world where coal preparation plants were generally smaller than in
Australia, mechanical cells have continued to play a predominant role in coal
flotation.
F L O TAT I O N F U N D A M E N TA L S
In the 1970s, flotation circuits typically comprised large numbers of cells, each
with their own motor and complex arrangements of paddles, and distribution
and froth collection systems. There have been significant developments in
flotation cell design in the intervening years. The introduction of round tank
designs in the 1980s heralded an almost exponential increase in individual tank
volumes. What may once have required literally hundreds of flotation cells can
now be achieved in a bank of four or five cells. Improved control systems and
the power reductions resulting from these economies of scale have eliminated
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
214 Beneficiation Technologies
the perceived problems associated with mechanical flotation in the latter half
of the 20th century. The development of mechanical flotation cells has main-
tained the basic process needs of the flotation process, those of residence time
and flotation kinetics.
Residence Time
Mechanical flotation cells are primarily designed on the concept of residence
time in the flotation cell or bank of cells. Fundamental circuit design requires
the calculation of the total flotation volume required to give sufficient time and
opportunity for the physical and chemical processes of flotation such as activa-
tion and contact times to occur. Residence time is calculated from the flow into
the cell and the effective cell volume. The required residence time is usually
determined by test work from either bench-scale or pilot-plant test work. If
laboratory-scale test work has been performed, the operational residence time
is determined using a scale-up factor. Historically, a scale-up factor of 2 has
been used from laboratory work. In mineral applications, rougher/scavenger
circuits may require residence times from 20 to 30 minutes, whereas cleaning
applications will be lower at 5 to 15 minutes residence times. In coal flotation,
residence times of 5 minutes are typically used.
Hydrodynamics of Flotation
Essentially, froth flotation depends on the combination of several prob-
abilities. The probability of a successful froth recovery is the product of these
probabilities:
• Pc = probability of collision of particles and bubbles
• Pa = probability of adhesion of particle to bubble
• Pd = probability of detachment in the turbulent region of the flota-
tion zone
• Pl = probability of levitation and recovery of values in the concentrate
Collision
The probability of a particle colliding with a bubble depends on the relative
sizes of particles and bubbles, hydrodynamics of the flotation environment,
but not on hydrophobicity. The probability increases as particle size increases.
Adhesion
The probability of a particle attaching to one or more bubbles is largely a
function on particle hydrophobicity. The probability decreases as particle size
increases.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Mechanical Cells for Fine Coal Flotation 215
Detachment
The probability of the particle/bubble aggregate persisting depends on hydro-
phobicity and the hydrodynamics of the system. The probability increases as
particle size increases.
Levitation
The probability of the particle reaching the concentrate overflow is the final
stage. If a stable particle/bubble aggregate forms, the aggregate must pass
through the pulp/froth interface to the overflow launder. This part of the pro-
cess requires a quiescent zone but also sufficient energy for froth movement
toward the discharge. Tao (2004) has provided an excellent summary of these
processes with particular reference to coarse and fine particles.
Number of Cells
To prevent the effects of short-circuiting, it is necessary to divide the total
volume required into a number of individual cells. Depending on the mineral
and the duty, the minimum number of cells varies from one to six. For coal
applications, a minimum of four cells is recommended. In a bank of flotation
cells, it is reasonable to assume that the fast-floating material will float first.
Based on the probability of a particle floating and the competition for floating/
transfer across the pulp/froth interface, these fast-floating particles will be the
optimum-sized particles (neither too coarse nor too fine) and the strongly
hydrophobic particles. In subsequent cells, the competition is reduced and
slower floating particles are given an increased opportunity to float. From a
purely statistical point of view, the advantages of multiple cells is clear. It is the
practical equivalent of successive points on the release analysis/ultimate flota-
tion curve. Using the traditional hog trough design for illustration and from an
idealized point of view, it might be imagined that the first cell could recover
about 50% of the liberated coal. The second cell might recover 50% of the
remainder (or 25%), and so forth, to give an overall recovery that approaches
100%. Clearly, this is somewhat stylized but serves to illustrate the principle of
multiple stages of flotation.
The actual residence time, within the bubble/particle contact region or
flotation zone, is on the order of seconds, and in a mechanical cell, particles
will pass through the contact zone several times before it exits the vessel. In the
absence of recycle, a column or Jameson cell allows a single pass through the
flotation zone.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
216 Beneficiation Technologies
Flotation Circuits
It is well known that a single flotation device cannot be optimized for both
grade (ash) and recovery. Flotation cells are classified according to the duty. The
first stage is referred to as roughing. Here, the aim is to produce a concentrate
and tailings with maximum recovery at an acceptable grade. Scavenger cells aim
to recover valuable material displaced to tailings.
Scavenger cells will typically be used to refloat the rougher tailings and
“scavenge” any values from these tailings. A bank of flotation cells will often
perform roughing and scavenging in the same bank with the first cells effec-
tively roughing and the last two or three cells scavenging the tailings. In coal
flotation, it is usual for the easy floating coal to be recovered in the first cells
with the more difficult to float coal, slow floating, and coarse particles in the
last cells.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Mechanical Cells for Fine Coal Flotation 217
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
218 Beneficiation Technologies
Euston (2010a) provided a review of multiple circuits for coal and com-
ments that the concept has been revisited every 10 years since the original pub-
lication by Firth. Arnold (1999) describes the concept of a rougher-scavenger
circuit for coal as a “grab and run” concept. In the “grab” stage, gentle flota-
tion conditions are applied (low air rate, low impeller speed, low addition of
reagents) to recover the best coal. This stage focuses on grade or ash and might
be used to produce a coking coal product in a mixed plant. The second stage
of scavenging (“run”) is aimed at recovery with less stress on product grade or
ash. Flotation parameters such as air rate, froth depth, and so on, are optimized
for recovery rather than grade. The paper discusses the concepts of cleaning the
scavenger product and a classification stage prior to scavenging to discard the
high-ash ultrafines, both interesting concepts but a little beyond the scope of
the present paper. Continuing the 10-year tradition, more details can be found
in Euston (2010b).
To summarize:
• Effective flotation requires a consideration of the time required for
particles and bubbles to have sufficient multiple interactions for
attachment.
• The physiochemical processes of attachment to occur are based on the
product of several probabilities.
• Flotation conditions for coarser and finer particles are quite different.
• Optimum conditions for recovery and grade (ash) cannot be achieved
in the same unit operation.
I N T E R N AT I O N A L P E R S P E C T I V E S A N D D E V E L O P M E N T S
As discussed previously, the decision to install flotation for fine coal depends on
both economic and technological factors, with an element of corporate or even
personal preference about whether or not the value to be received is worth the
capital and ongoing operating costs. The decision on the choice of technology is
also based on corporate or individual preferences in addition to the coal quality
and the product specifications. In general, coals with high clay content in the
fines or if lower ash is required in the fines product may benefit from column-
type flotation, although this will normally be at the expense of recovery.
United States
In the Unites States, there is an increasing trend toward mechanical cells,
although, as mentioned previously, corporate preference plays a significant role.
Traditional square or “hog trough” designs dominate the market. These Wemco
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Mechanical Cells for Fine Coal Flotation 219
machines are self-aspirated and as such are less operator intensive and produce
more consistent results without constant attention. Cell sizes are available up to
42.5 m3 (1,500 ft3) with a size range that allows for a bank of four to six cells in
a compact and efficient circuit. The majority of the cells in operation are 8.5–28
m3 (300–1,000 ft3). A typical flotation circuit will consist of one or more rows
of cells, with each row consisting of four to five cells in series. In general, a pulp
residence time within the separation vessel is in the range of 3–5 minutes. A
bank of cells can be seen in Figure 1.
For plants with a high ash or clay content, flotation columns are often used.
Recovery can often suffer, however, and both high recovery and low ash can be
achieved with a rougher-cleaner arrangement.
United Kingdom/Europe
The coal industry in Eastern Europe has declined in recent years. Coal flotation
plants in Eastern Europe tend to follow the U.S. philosophy of small, compact
banks of mechanical flotation cells with the focus on consistent operation with
minimum operator involvement. Figure 2 shows two typical banks of compact
14.2-m3 (500-ft3) Wemco flotation cells operating in South Wales.
Another popular concept in Europe is the self-contained flotation/filtra-
tion plant (see Figure 3). Dewatering of fine coal concentrate is best performed
with vacuum filtration. Horizontal belt filters are the industry preference, as
they provide consistent operation over a wide range of feed variability, both in
terms of quality and metric tons per hour.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
220 Beneficiation Technologies
South Africa
In 2002, Opperman et al. (2002) reported that the Witbank coals of South
Africa were difficult to float with low yields essentially making the process
uneconomic. High energy and high reagent costs were the major sources of
problems. More recent private communications indicate that this attitude pre-
vails. Much South African coal is for power station use and the fines are often
used without processing.
Opperman et al. (2002) describe the Anglo Coal experience at Goede-
hoop colliery and their investigations of coal flotation options and results. The
authors compared the flotation response of Jameson and Wemco pilot test
units (see Table 3), concluding that although the mechanical cell gave the best
recovery and ash combination, both technologies had benefits. The Jameson
cell did not appear to provide sufficient energy in the downcomer to sufficiently
absorb the frother. The article describes in detail the development of the hybrid
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Mechanical Cells for Fine Coal Flotation 221
multicell. The results are summarized in Table 3. The multicell has been further
refined as the dual cell, and a number of installations are currently being com-
missioned in South Africa. The dual-cell concept uses two cells per bank in a
rougher-scavenger combination.
Australia
Mechanical Flotation Stands the Test of Time
During the growth in the Australian coal industry in the 1980s, many of the
new, large coal preparation plants installed mechanical flotation technolo-
gies. These new generation plants with their high throughputs required large
numbers of mechanical cells. Figure 4 shows six banks of 14.2-m3 (500-ft3)
self-aspirated cells (then the largest available) installed at German Creek in
Queensland, Australia, in 1982. This installation has just been entirely refur-
bished with complete new mechanisms. Figure 4 shows the recently upgraded
flotation circuit at German Creek.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
222 Beneficiation Technologies
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Mechanical Cells for Fine Coal Flotation 223
90.0
60.0
50.0
Yield, %
40.0
30.0
20.0
10.0
0.0
1 2 3 BRN Coke + Duralie Duralie
Dirty Thermal Coke
Roseville
Coal Type
Figure 6 shows the buildup of fine coal in the tailings thickener when not using
secondary flotation. Figure 7 shows the improved water clarity of the tailings
thickener when using secondary flotation.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
224 Beneficiation Technologies
The first three cells, shown in Figure 8 during construction, are used as rough-
ers. The other two cells can be configured as rougher or cleaner cells depending
on coal type. At the time of this writing, the cleaners are about to be commis-
sioned. The inherent flexibility of this circuit and the innovative approach of
the client will allow extensive on site research to be carried out.
China
China is currently the world’s largest coal producer with limited exports. A
search of flotation technologies in China reveals many local manufacturers, and
it is likely that these will dominate the local industry. Several types of flotation
cells are shown in Figure 9.
CONCLUSIONS
Froth flotation is currently the only economic and practical process for the
treatment of the finest coal fractions. In principle, froth flotation is a simple
concept, the attachment of bubbles to receptive particles and their subsequent
removal from the flotation device as a froth concentrate. In practice, however,
the process is technically complex, relying on a number of physicochemical and
hydrodynamic disciplines.
The need for residence time and the probabilities associated with the
recovery of values are best achieved in multiple reaction vessels and ideally
using separate circuits to achieve product grade and recovery. These require-
ments are well satisfied by the operation in traditional mechanical flotation
cells. The economics of treating the finest coal are often tenuous, and the use
of complex circuits is often not attractive, both financially and culturally. The
introduction of single-stage column and similar technologies to the coal indus-
try in the 1980s resulted in a rapid acceptance of these new technologies as they
promised simplicity and minimal operator involvement, albeit at the expense
of coal recovery.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Mechanical Cells for Fine Coal Flotation 225
Figure 8 Rougher flotation cells being installed at Rix’s Creek, Hunter Valley, New
South Wales, Australia
REFERENCES
Arnold, B.J. 2000. The “grab and run” revisited—Improving selectivity between
organic and inorganic components in conventional coal flotation. Int. J. Miner.
Process. 58:119–128.
Crisafulli, P., and James, T. 2008. Secondary flotation at Stratford CHPP—A mechani-
cal cell processing Jameson cell tailings. In Proceedings of the 12th Australian Coal
Preparation Society Conference, Sydney. pp. 382–390.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
226 Beneficiation Technologies
Euston, J.A. 2010a. Two stage coal flotation using a mechanical cell. In XVI Interna-
tional Coal Preparation Congress, Proceedings. Edited by R. Honaker. Littleton,
CO: SME. pp. 382–390.
Euston, J.A. 2010b. Mechanical flotation cells in coal preparation—a technical review
and international perspective. In 13th International Coal Preparation Conference,
Mackay, Australia. pp. 339–351.
Firth, B.A., Swanson, A.R., and Nicol, S.K. 1979. Flotation circuits for poorly floating
coals. Int. J. Miner. Process. 5:321–334.
Heiser, N. 1996. Coal Flotation Technical Review—A Report on an ACARP Project.
Undertaken by Mike Williamson and Joe Sanders, The Australian Coal Review,
October.
Lynch, A.L., Harbort, G., and Nelson, M. 2010. Coal flotation. In History of Flotation.
Carlton, VIC: Australasian Institute of Mining and Metallurgy.
Opperman, S.N., Nebbe, D., and Power, D. 2002. Flotation at Goedehoop colliery. J. S.
Afr. Inst. Mining Metall. (October):405–410.
Sanders, G.J., and Williamson, G.J. 1996. Coal Flotation Technical Review C4047—An
ACARP Commissioned Study. February.
Tao, D. 2004. Role of bubble size in flotation of coarse and fine particles—A review.
Sep. Sci. Technol. 39(4):741–760.
Wade, W. 2008. Float cell breakthrough. Aust. Min. 10 March.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Gravity Separators for
Ultrafine Coal Cleaning
Rick Honaker
ABSTRACT
The application of density-based separators for upgrading ultrafine coal has been
studied and utilized in select applications. In some cases, conventional separators
such as the spiral concentrator have been commercially used to provide an effective
density-based separation for particles finer than 0.15 mm. Relative to the spiral
application, a higher level of efficiency and capacity may be realized by continuous
enhanced-gravity separators, which utilize fundamental density-based separation
principles to achieve ultrafine particle concentration in a centrifugal field. This
chapter reviews the processes available to achieve effective gravity separations for
ultrafine coal particles and their associated separation performances.
INTRODUCTION
The upgrading of coal below a particle size of 0.15 mm (100 mesh) is predomi-
nantly achieved using froth flotation in operating preparation plants world-
wide. The flotation process separates coal from the associated rock particles by
exploiting the natural surface hydrophobicity of the coal particles. However,
some coal sources are not effectively treated by froth flotation due to surface
oxidation. In these cases, collectors or other surface-modifying chemicals may
be applied to selectively generate a hydrophobic surface on the coal, which may
add significant operating cost. Another option is to utilize a density-based sepa-
rator with the capabilities to achieve an effective treatment of ultrafine coal.
Density-based separators may also be needed to separate pyrite from coal
prior to flotation. Several studies have reported that coal pyrite surfaces are
hydrophobic under certain flotation conditions, thereby reducing the effec-
tiveness of the flotation process in reducing the total sulfur content (Hurt
and Aplan 1991; Yoon et al. 1991). In fact, a few plant operators treating coal
containing moderate-to-high levels of pyritic sulfur choose to avoid flotation
227
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
228 Beneficiation Technologies
S P I R A L C O N C E N T R AT O R
The separator most commonly used for cleaning the 1 × 0.15 mm particle
size fraction is the spiral concentrator because of its operational simplicity
and relatively low capital and operating cost. Separation efficiency has been
significantly improved as a result of focused research efforts that led to the
development of the compound spiral. The typical spiral concentrator receives
feed from the underflow of a bank of raw coal classifying cyclones after being
diluted to a desired solids concentration of around 30% by weight. With mass
solids feed rates of around 2.5 tph per spiral start, highly effective specific grav-
ity cutpoints as low as 1.70 can be realized for the majority of the particles in
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Gravity Separators for Ultrafine Coal Cleaning 229
–0.15 mm
–1 mm
Feed
Classifying
Cyclone
1 × 0.15 mm
Tailings
Sieve-bend Screen-bowl
Product
–0.15 mm
Thickener
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
230 Beneficiation Technologies
Table 1 Process performance and efficiency data achieved from cleaning 0.21 ×
0.044 mm Coalberg seam coal under optimum conditions in a spiral concentrator
Efficiency Parameters Test 1 Test 2 Test 3 Test 4 Test 5
Feed ash, % 31.85 36.82 35.26 36.04 38.45
Product ash, % 8.04 11.71 13.30 16.14 19.73
Tailings ash, % 69.10 74.92 85.61 90.51 91.10
Mass yield, % 50.19 60.28 69.63 73.24 83.02
Combustible recovery, % 74.99 84.23 93.25 96.03 97.78
Ash rejection, % 89.50 80.83 73.73 67.21 48.58
Separation efficiency, % 64.49 65.06 66.98 63.24 46.36
Separation density, d50 1.65 1.89 2.06 2.17 2.25
Probable error, Ep 0.22 0.23 0.22 0.13 0.13
High-density bypass, % 5.03 15.05 19.95 27.55 42.36
Low-density bypass, % 16.93 9.77 2.00 0 0
Organic efficiency, % 78.11 86.88 96.13 98.49 99.78
significantly improve performance. The samples taken from all process streams
were screened to remove the particles smaller than 0.044 mm. As such, the data
in Table 1 represent the assays obtained from the 0.21 × 0.044 mm (65 × 325
mesh) particle size fraction.
Product ash values ranging from 8.04% to 19.73% were achieved from 0.21
× 0.044 mm Coalberg seam coal. The ash value of the feed coal was greater than
30%. To achieve the lower product ash values, a substantial amount of low-
density material was bypassed to reject, which resulted in relatively low organic
efficiency values. Likewise, a large amount of high-density, high-ash value par-
ticles were bypassed to the product when targeting relatively high product ash
values. The most desirable performance corresponded to a product ash value of
13.30%, which provided an organic efficiency of 96%.
To achieve the desired performance, the mass throughput capacity was
required to be reduced to about 0.5 tph/start from a typical value of about
3 tph/start. This finding reflects (1) the impacts of increased particle popu-
lation with a reduction in particle size and (2) the need for greater particle
retention time. Similar findings were reported by Benusa and Klima (2008),
although the stated impacts of feed solids and volumetric flow were less signifi-
cant. The optimum volumetric feed rate was approximately 50 L/m (13 gpm)
at a feed solids concentration of 15% by weight.
An in-plant study was performed to evaluate the separation performance
of an SX7 Multotec spiral when applied to treat the underflow of a second-
ary 15-cm-diameter classifying cyclone at a preparation plant treating Illinois
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Gravity Separators for Ultrafine Coal Cleaning 231
No. 6 seam coal (USA). The nominal 0.15 × 0.044 mm spiral feed contained
about 12.0% solids by weight. The volumetric and solid mass feed flow rates
were 70 L/m (18 gpm) and 0.6 tph, respectively. As shown in Table 2, signifi-
cant ash and total sulfur reductions were achieved for all size fractions in the
feed. For the +0.044 mm coal, 60% of the ash-forming minerals were rejected
while 48.3% rejection of ash was obtained for the overall coal. Likewise, 47.6%
total sulfur rejection was achieved. To achieve an acceptable final product, the
0.044 mm fraction will need to be removed by classification or screening.
C E N T R I F U G A L F L O W I N G - F I L M C O N C E N T R AT O R S
Particles within a thin film of water moving across a solid surface segregate
according to density due to a differential streamline velocity profile. High-
density particles settle onto the solid surface, whereas light particles remain in
the high-velocity streamlines located a distance away from the surface. Spiral
concentrators and riffle tables are examples of flowing-film separators that are
commonly used for particles coarser than 0.15 mm. Continuous centrifugal
concentrators utilizing the flowing-film principle have been developed to allow
density-based separations for ultrafine particles. Commercial units include the
Falcon concentrator and the MGS.
Falcon Concentrator
The essential feature of a continuous Falcon concentrator is a vertically aligned,
open-topped cylindrical bowl that is mounted on a rotating shaft, as shown in
Figure 2. A centrifugal force up to 300 g’s can be produced to cause deposition
and stratification of the fine particles against the inside of a smooth centrifugal
bowl. The coal slurry is continuously introduced at the bottom of the spinning
bowl by means of a conduit extending downward along the axis of rotation. The
feed slurry is then impelled to the wall of the bowl by an impeller, which causes
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
232 Beneficiation Technologies
Feed
Rinse Water
Heavies Heavies
Dilution Dilution
Water Water
Lights Outlet
Heavies Heavies
Outlet Outlet
Rotor Bowl
Cleanout
Outlet
stratification along the inclined lower section of the bowl, called the migra-
tion zone, due to differential acceleration. In this zone, the enhanced-gravity
field is resolved into two force components. The strong component normal to
the wall is the concentrating gravity field that provides the strong g-forces for
the hindered-settling processes and density stratification of the particles. The
weak driving component parallel to the bowl wall pushes the stratified solids
up toward the top of the bowl. An overflow lip that has an internal diameter
less than the bowl diameter restricts the particle bed from reporting to the
overflow. The combination of the lack of a vertical force component and the
presence of an overflow lip causes the heavy pyrite and ash-bearing particles to
come to rest while the centrifugal force assists the heavy particles into a slot that
exists around the circumference of the bowl. Mass transport chutes and pinch
valve-nozzle assemblies placed at equal distances in the slot allow discharge of
the heavy particles into an underflow launder. At the same time, light particles
forming the particle bed furthest from the bowl wall move upward and over the
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Gravity Separators for Ultrafine Coal Cleaning 233
Table 3 Total sulfur reductions achieved on a particle size-by-size basis from the
treatment of –0.60 mm Pittsburgh No. 8 coal by C40 Falcon concentrator
Total Sulfur, %
Particle Size Combustible
Fraction, mm Feed Product Tails Recovery, %
+0.60 3.21 1.92 12.30 90.8
0.60 × 0.30 3.32 1.67 12.60 92.7
0.30 × 0.212 4.26 1.71 9.92 92.9
0.212 × 0.150 3.06 1.77 13.70 93.6
0.150 × 0.075 3.37 1.77 22.20 93.7
0.075 × 0.037 6.52 3.12 20.50 94.4
overflow lip of the bowl. These light particles report as final product with other
particles (i.e., colloidal clay) that are too fine to be affected by the enhanced
gravitational force.
Water-Only Separations
The ability of the Falcon concentrator to effectively clean coal to a particle
size as small as 0.037 mm has been extensively studied using a 25-cm-diameter
pilot-scale unit (Honaker et al. 1996) and a 1-meter-diameter full-scale unit
(Honaker 1999). Performance results from the treatment of a relatively large
number of coal samples indicate an exceptional ability to reduce sulfur content
in the ultrafine fractions while also achieving effective ash reductions. Using
the full-scale unit at a mass flow rate of 78 tph, nearly 75% of the ash-bearing
material and 60% of the sulfur from a Pittsburgh No. 8 seam coal were removed
while recovering 80% of the mass yield to the product. Table 3 shows the total
sulfur reductions realized on a particle size-by-size basis. Organic efficiency
(i.e., ratio of actual recovery over theoretical recovery) was reported to be about
96%.
Dense-Medium Separations
Research findings reported by Honaker and Patil (2001) indicate that the
separation performances provided by an enhanced-gravity concentrator can
be significantly enhanced using a dense medium. The medium was formed
using magnetite in which 90% of the material was smaller than 0.017 mm.
A test program was performed in which the operating parameter values were
significantly varied to obtain the data needed to develop the empirical models
used to optimize their values. The optimized performances shown in Figure 3
were obtained by varying the medium specific gravity from 1.35, to realize the
lowest product ash value, to 1.50 to produce the highest product ash value. The
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
234 Beneficiation Technologies
100
80
60
Clean Coal Yield, %
40
20
Washability
Dense-Medium Falcon
Optimized Dense-Medium Falcon
0
0 5 10 15 20 25
Product Ash, %
relatively small difference between the mass yield values from the optimized
tests and the theoretical values represented by the washability curve indicates a
very high level of efficiency over the entire range of product ash values.
C E N T R I F U G A L F L U I D I Z E D - B E D C O N C E N T R AT O R
The use of fluidized-bed units for achieving effective density-based separations
has historically been a common practice within the coal industry. The Chance
cone is an example in which an upward current of water is used to suspend sand
particles and create a medium through which the high-density, high-ash parti-
cles settle. Within the last few decades, fluidized-bed separators, which are also
known as teeter beds, have been used as an alternative to spiral concentrators
for upgrading coal within the particle size range of 1 × 0.15 mm. In this applica-
tion, the fluidization water is used to suspend high-density, high-ash particles
entering in the feed and, as a result, create an autogenous medium. As a result,
separation efficiencies achieved over a particle size range of around 4:1 are
exceptional. However, performances deteriorate significantly with a reduction
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Gravity Separators for Ultrafine Coal Cleaning 235
Feed
Overflow Overflow
Product Product
Underflow Underflow
Tailings Tailings
Elutriation Water
Knelson Concentrator
The Knelson continuous variable discharge (CVD) concentrator employs flu-
idized particle bed separation concepts in a mechanically applied centrifugal
field. As shown in Figure 4, the Knelson concentrator operates by introducing
water through a series of fluidization holes located in the rings that circle the
circumference of a bowl. The bowl, which has a truncated cone shape, is rotated
at speeds that provide a centrifugal field up to 200 times gravity. Feed slurry is
introduced through a tube that directs the material toward the bottom cen-
ter. Upon reaching the bottom, the slurry is driven outward and up the cone
wall toward the rings. The fluidization water entering in the rings provides an
inward velocity that allows the creation of a fluidized particle bed comprised of
heavy particles. The high-density particles that pass through the fluidized par-
ticle bed are extracted through a series of controlled pinch valves located along
the circumference in the center of a concentration ring. The discharge ports are
designed to handle mass flows equivalent to 1%–50% of the feed.
A pilot-scale unit of the Knelson CVD concentrator with a maximum
throughput capacity of 5 tph was evaluated in an operating coal preparation
plant treating both steam and metallurgical coal sources (Honaker and Das
2004; Honaker et al. 2005). The feed to the unit was provided from the under-
flow of a bank of 15-cm-diameter desliming classifying cyclones. Approxi-
mately 40% of the coal had a particle size between 0.15 mm and 0.044 mm and
an equal amount was finer than 0.044 mm. In an effort to improve recovery, air
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
236 Beneficiation Technologies
100
90
80
70
60
Recovery, %
50
40
30 Coalberg Coal
20 Washability
Experimental
10 Release
0
0 4 8 12 16 20 24
Product Ash, %
bubbles were injected into the feed stream to generate particle–bubble aggre-
gates that have a lower density than the individual coal particles.
The performance data shown in Figure 5 is an example of the ability to
enhance the separation performance of an ultrafine coal using EGS. The steam
coal (Coalberg seam, United States) was a poorly floating material and the
selectivity achieved by the flotation process was poor, as indicated by the release
data. The separation performances achieved by the Knelson CVD unit were
superior to the froth flotation performance represented by the release curve.
The optimum separation performances shown in Table 4 indicate that the
Knelson unit reduced the ash value from 17.8% to 10.1% while recovering 81%
of the combustibles. Injecting air bubbles into the feed increased the recovery
by 13 absolute percentage points. The relative separation specific gravity values
varied from 1.3 to 1.8 with corresponding probable error values around 0.20.
Significant bypass (low- and/or high-density particles) occurred in all tests.
CENTRIFUGAL JIG
Commercially available centrifugal jig units include the Kelsey and Altair jigs.
The Altair jig consists of a rotating bowl, which is placed inside a static casing
having separate launders for collecting the concentrate and tailing samples
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Gravity Separators for Ultrafine Coal Cleaning 237
(Figure 6). The rotating bowl contains a cylindrical screen with a lip, whose
height can be adjusted to vary the natural depth of the ragging bed, which
remains in a vertical position on the screen due to bowl rotation. The feed
slurry, which is introduced from the top at the center of the rotating bowl, is
distributed into the ragging bed on the screen by the diffuser plate placed under
the feed inlet. Pressurized water is periodically injected under the bed through
four pulse-blocks to cause alternating dilation and contraction of the ragging
and feed bed. This, coupled with the high centrifugal force generated by the
rotation of the bowl, provides a hindered-settling environment required for
the jigging to occur, which results in physical separation of particles of varying
density. The tailings material settles through the ragging bed and screen into
the hutch and reports to the tailings launder through the discharge ports. The
clean coal particles do not settle through screen and, thus, report to the con-
centrate launder.
Tests conducted on Illinois No. 5 coal containing particles finer than 1 mm
reduced the ash value from 28.00% to 6.84% while achieving a recovery of 82%
(Mohanty et al. 1999, 2002). Ash and total sulfur reduction averaged about
85% and 42%, respectively. The separation performances shown in Figure 7
indicate that tests performed with no ragging provided inferior results. When
granite ragging material was used, performance substantially improved. The
results indicate some loss of coal to the tailings stream. The specific gravity cut-
point ranged between 1.43 and 1.54 with probable error values varing between
0.08 and 0.17.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
238 Beneficiation Technologies
100
80
Combustible Recovery, %
60
40
20
Feed Washability
Ragging Tests
No Ragging Tests
0
0 20 40 60 80 100
Ash Rejection, %
CONCLUSIONS
General applications of density-based separations for treating ultrafine coal
(i.e., particles finer than 0.15 mm) include coal sources that have difficult-
to-float characteristics and those that contain a significant amount of pyritic
sulfur. Spiral concentrators have been evaluated for these applications and
implemented into a few operating preparation plants. Available plant data
indicate the ability to significantly reduce the ash value and total sulfur content.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Gravity Separators for Ultrafine Coal Cleaning 239
REFERENCES
Benusa, M.D., and Klima, M.S. 2008. An evaluation of a two-stage spiral processing
ultrafine bituminous coal. Int. J. Coal Prep. Util. 28(4):237–260.
Hirt, W.C., and Aplan, F.F. 1991. The influence of operating factors on coal recovery
and pyritic sulfur rejection during coal flotation. In Processing and Utilization of
High Sulfur Coals. 4th ed. Edited by P.R. Dugan, D.R. Quigley, and Y.A. Attia.
Amsterdam: Elsevier Science Publishers. pp. 339–356.
Honaker, R.Q. 1998. High capacity fine coal cleaning using an enhanced gravity con-
centrator. Miner. Eng. 11(12):1191–1199.
Honaker, R.H., and Das, A. 2004. Ultrafine coal cleaning using a centrifugal fluidized-
bed separator. Coal Prep. 24(1-2):1–18.
Honaker, R.Q., and Patil, D.P. 2001. Parametric evaluation of a dense-medium process
using an enhanced gravity concentrator. Coal Prep. 22(1):1–17.
Honaker, R.Q., Wang, D., and Ho, K. 1996. Application of the Falcon concentrator for
fine coal cleaning. Miner. Eng. 9(11):1143–1156.
Honaker, R.Q., Das, A., and Nombe, M. 2005. Improving the separation efficiency of
the Knelson concentrator using air injection. Coal Prep. 25(2):99–116.
Honaker, R.Q., Jain, M., and Saracoglu, M. 2007. Ultrafine coal cleaning using spiral
concentrators. Miner. Eng. 20(14):1315–1319.
Majumder, A.K., Bhoi, K.S., and Barnwal, J.P. 2007. Multi-gravity separator: An
alternative gravity concentrator to process coal fines. Miner. Metallurg. Process.
24(3):133–138.
Menéndez, M., Gent, M., Torano, J., and Diego, I. 2007. Optimization of multigravity
separation for recovery of ultrafine coal. Miner. Metallurg. Process. 24(4):253–263.
Mohanty, M.K., and Honaker, R.Q. 1999. Evaluation of the Altair centrifugal jig for
fine particle separations. Coal Prep. 20(1):85–106.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
240 Beneficiation Technologies
Mohanty, M.K., Honaker, R.Q., and Patwardhan, A. 2002. In-plant evaluation of the
Altair centrifugal jig for fine coal cleaning. Miner. Eng. 15:157–166.
Paul, B.C., and Honaker, R.Q. 1993. Production of Illinois Basin Compliance Coal Using
Enhanced Gravity Separation. Final Technical Report. Report Number 93-1/5.1B-
1P. Illinois Clean Coal Institute.
Richards, R.G., MacHunter, D.M., Gates, P.J., and Palmer, M.K. 2000. Gravity
separation of ultra-fine (–0.1 mm) minerals using spiral separators. Miner. Eng.
13(1):65–77.
Riley, D.M., and Firth, B.A. 1993. Application of an enhanced gravity separator for
cleaning fine coal. In Proceedings, 10th International Coal Preparation Conference,
Lexington, KY. pp. 46–65.
Venkatraman, P., Luttrell, G.H., and Yoon, R.H. 1995. Fine coal cleaning using the
multi-gravity separator. In Proceedings of the High Efficiency Coal Preparation:
An International Symposium. Edited by S.K. Kawatra. Littleton, CO: SME. pp.
109–117.
Yoon, R.H., Lagno, M., and Luttrell, G.H. 1991. On the hydrophobicity of coal pyrite.
In Processing and Utilization of High Sulfur Coals. 4th ed. Edited by P.R. Dugan,
D.R. Quigley, and Y.A. Attia. Amsterdam: Elsevier Science Publishers. p. 241.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Design and Operating
Guidelines for Combined
Water-Only Cyclone and Spiral
Circuits
Gerald Luttrell, Zulfiqar Ali, Andy Dynys,
Larry Watters, and Robert Moorhead
ABSTRACT
The methods used to treat coarse and fine coal streams in modern coal preparation
plants have become largely standardized. For optimum efficiency, coal particles
larger than about 1 mm are upgraded using dense-medium separators, whereas
particles smaller than about 0.15 mm (100 mesh) are cleaned using froth flota-
tion. No single standard method has yet emerged as the most efficient option for
treating run-of-mine feeds in the size range between 1 mm and 0.15 mm. From
a historical perspective, gravity-based separators such as water-only cyclones have
been a popular commercial choice for treating this intermediate size fraction. In
recent years, however, the emergence of a new generation of two-stage compound
spirals has begun to largely displace water-only cyclones, generally because of the
higher efficiency offered by this multi-stage cleaning device. Despite this fact, several
recent studies have suggested that other gravity-based separators, such as water-
only cyclones, can be effectively incorporated into multi-stage circuits to improve
capacity, lower costs, and improve separation efficiency. This chapter discusses
application guidelines for configuring such circuitry and offers recommendations
for optimizing overall circuit performance.
INTRODUCTION
Modern coal preparation facilities incorporate a wide variety of solid–solid
separation processes for coal upgrading. Dense-medium processes, which
241
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
242 Beneficiation Technologies
DM Vessel
Coarse Jig
DM Cyclones
WOC
Spirals
HydroFloat
Reflux Classifier
Teeter Bed
Fine Spirals
Froth Flotation
Figure 1 Approximate effective size ranges for different coal cleaning processes
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Combined Water-Only Cyclone and Spiral Circuits 243
biases also appear to contribute to the large variations that are observed in how
fine coal is cleaned.
Two of the most popular methods for treating fine coal in the 1 × 0.15 mm
size range are water-only cyclones (WOCs) and spirals. A WOC is similar to
a classifying cyclone, but typically has a stubby, wide-angled conical bottom
(Figure 2). Separations occur in a WOC because of differences in the settling
rates of coal and rock in the centrifugal field within the cyclone. The separa-
tion does not utilize any external medium such as magnetite but is enhanced
by the formation of autogenous medium created by the natural fines already
in the feed slurry. WOCs typically utilize a truncated cone bottom and wider
included angle (usually 60–120°) that promotes the formation of a refuse bed
which prevents lighter coal particles from reporting to underflow. WOCs are
also typically operated with a relatively long vortex finder that “vacuums up”
the light coal particles off the top of the bed of autogenous medium. These
units are often employed in two stages or in combination with other water-
based separators to improve performance.
A spiral consists of a corkscrew-shaped conduit with a modified semicir-
cular cross section (Figure 3). During operation, feed slurry is introduced to
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
244 Beneficiation Technologies
the top of the spiral and is permitted to flow by gravity along the helical path
to the bottom of the unit. Particles in the flowing film are stratified such that
lighter coal particles are forced to the outer wall of the spiral, whereas heavier
particles are forced inward to the center of the spiral. The segregated bands of
heavy to light materials are collected at the bottom of the spiral. Adjustable
diverters (called splitters) are used to control the proportions of particles that
report to the various products. A three-product split is usually produced, giv-
ing rise to three primary products containing clean coal product, refuse, and
misplaced “middlings.” Because of the low unit capacity, spirals are usually
arranged in groups or banks that are fed by an overhead radial distributor. To
save space, several spirals (two or three) may be intertwined along a single cen-
tral axis. Modern coal spirals typically incorporate two stages of cleaning along
a single support column to reduce misplacement. Typically, the first four (or
three) spiral turns are used to produce a throw-away reject product, after which
the clean coal and middlings products are remixed in a pulping box and then
re-cleaned using three (or four) additional spiral turns. The products from the
second stage include final clean coal, final reject, and middlings that can be sent
to clean coal, reject, or recycled back to be reprocessed. In most cases, recycling
of middlings is generally recommended to improve the overall sharpness of the
separation (Luttrell et al. 1998; Bethell and Arnold 2003).
O P E R AT I N G A N D D E S I G N PA R A M E T E R S
Water-Only Cyclones
The basic working features of a WOC are relatively simple, but numerous design
and operating variables can influence its performance. The most imporatant of
these variables are (1) the length of the vortex finder, (2) the included angle of
the truncated cone and apex, and (3) the apex-orifice diameter. Increasing the
length of the vortex finder extends the entry of the vortex finder into a region of
higher-density particles, rotating in the lower region of the cyclone. As such, a
longer vortex finder increases the specific gravity of separation (SG50). Increas-
ing the included angle of the truncated cone and/or the apex as well as decreas-
ing the apex-orifice diameter also increases the SG50 of the cyclone. All of these
adjustments increase the retention of particles in the autogenous zone of the
cyclone, thus making it more difficult for lower-density particles to penetrate
this zone, and, therefore, they are re-directed upward toward the vortex finder.
In addition, decreasing the apex diameter also increases the volumetric split
of the cyclone to the overflow. This results in greater drag forces, which pro-
mote directing a greater percentage of particles to the overflow. Generally, the
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Combined Water-Only Cyclone and Spiral Circuits 245
Low SG
Coal
High SG
Rock
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
246 Beneficiation Technologies
Spirals
Spirals are capable of maintaining good recoveries of clean coal when properly
operated (Luttrell et al. 2007). One of the most important operating variables
is feed rate. Ideally, spirals should be configured to operate at a dry solids feed
rate of about 2.3 t/h (2.5 ton/h) per spiral start (Li et al. 1993; Subasinghe
et al. 1991, 1992; Holland-Batt 1994). The SG cutpoint increases rapidly as
the feed tonnage exceeds this value, making it difficult to maintain an SG
cutpoint that would best optimize those in the plant dense-medium circuits.
The Ep value also generally diminishes as the tonnage rate increases. Spirals
must be provided with an adequate and stable slurry flow rate of approximately
7–8 m3/h (35–40 gpm) to work properly. Too little flow can result in sluggish
movement of solids along the interior surface of the spiral that can eventually
lead to beaching/sanding, whereas too much flow can cause high-density rock
to report inadvertently to the low-density stream. A well-designed and properly
maintained feed distributor is essential to ensure that each spiral receives the
sample feed rate to avoid different SG50 cutpoints that would adversely impact
performance. Typically, the slurry level in the distributor should be maintained
at 30–45 cm (12–18 in.) to maintain the head necessary to obtain consistent
flow rates from the discharge ports. When the proper flows have been set, spiral
performance can be fine-tuned using the splitter positions. The degree to which
splitter positions influence the SG50 cutpoint is highly dependent on particle
size and washability characteristics (Mikhail 1988; King et al. 1992).
S E PA R AT I O N P E R F O R M A N C E
A typical set of size-by-size partition curves for a bank of industrial WOCs
are shown in Figure 4. Since the WOC technology separates particles based
on differences in effective mass, the performance of the process is affected by
particle size as well as particle shape (Bull et al. 1987). As a result, the expected
density cutpoint (SG50) varies considerably for each size fraction treated by the
process (i.e., finer particles are separated at much higher densities than coarser
particles). Very fine particles have a low mass and tend to be entrained into the
overflow stream by drag forces, regardless of density. Very large particles have
a high mass and tend to preferentially report to the underflow, even if they are
of low density. This characteristic drift in SG cutpoint is plotted in Figure 5 for
several sets of industrial WOC circuits. The decline in SG50 with increasing
particle size is undesirable because it results in an overall partition curve that is
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Combined Water-Only Cyclone and Spiral Circuits 247
1
1.7 × 1.0 mm
1.0 × 0.6 mm
0.9 0.6 × 0.25 mm
0.25 × 0.15 mm
0.8
0.7
0.6
Partition Factor
0.5
0.4
0.3
0.2
0.1
0
1.2 1.4 1.6 1.8 2 2.2 2.4
Specific Gravity
substantially less efficient than that obtained for any of the individual size frac-
tions (Abbott 1981; Luttrell et al. 2000). This phenomenon forces operators
to select between operating conditions that either (1) throw away a significant
portion of low-density coal particles in order to produce an acceptable clean
coal ash in the finer size fractions or (2) recover the coarser coal particles and
tolerate a relatively high ash content in the finer size fractions. Although nei-
ther of these options is ideal, the WOC system does provide a variable cleaning
option when targeting relatively low-density SG cutpoints for particles in the
coarser size fractions. In fact, the Ep values obtained in the coarser size fractions
are typically quite low and provide relatively efficient separations (Schlepp and
Schmidt 1988).
Figure 6 provides a typical set of partition curves for a modern spiral sepa-
rator. Much like WOCs, fine particles having a low mass tend to be hydrauli-
cally carried by the fluid drag forces into the clean coal product regardless of
their density. Since the bulk of the water flow reports to clean coal, so too does
the vast majority of the fine particles. Consequently, the density cutpoint for
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
248 Beneficiation Technologies
2.7
WOC “A”
WOC “B”
WOC “C”
2.5
2.3
2.1
SG Cutpoint
1.9
1.7
1.5
1.3
0.1 1 10
Particle Size, mm
Figure 5 Effect of particle size on density cutpoint for three industrial installations
of water-only cyclones
spirals tends to increase as the particle size increases from the finest grain size
treated up to a critical particle diameter of approximately 0.4–0.6 mm. Above
this critical size, larger particles in the flowing film passing down the spiral
begin to experience body forces that act in direct proportion to their particle
size. These body forces effectively push the larger particles along in the flowing
liquid that reports to the clean coal product. Since the body forces depend on
size (and shape) and are independent of density, the density cutpoint for the
largest particles actually tends to increase as the particle size increases further.
The net effect of the combination of the settling and drag forces is a partition
curve that is minimum in the 0.4–0.6 mm size range and increases for larger
or smaller particles. A summary plot illustrating this phenomenon for serveral
industrial spiral operations is shown in Figure 7. The flattening of the SG50
versus particle size curve is actually desirable because it produces a composite
partition curve for the entire size range that is better than if a wider drift of
SG50 values had occurred. This inherent advantage is a major contributing
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Combined Water-Only Cyclone and Spiral Circuits 249
0.9
0.8
0.7
0.6
Partition Factor
0.5
0.4
0.3
0.2
1.7 × 1.0 mm
1.0 × 0.6 mm
0.1 0.6 × 0.25 mm
0.25 × 0.15 mm
0
1.2 1.4 1.6 1.8 2 2.2 2.4
Specific Gravity
COMBINED CIRCUITRY
One interesting approach for dealing with the inherent limitations of WOCs
and spiral separators is to integrate these two unit operations using multi-stage
circuitry. As early as two decades ago, Mikhail et al. (1988) demonstrated
experimentally that a combination of WOCs and spirals can help achieve
optimum recovery of coal depending on the washability of the fine particles.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
250 Beneficiation Technologies
2.5
Spiral “A”
Spiral “B”
Spiral “C”
2.3
2.1
SG Cutpoint
1.9
1.7
1.5
1.3
0.1 1 10
Particle Size, mm
Figure 7 Effect of particle size on density cutpoint for three industrial installations
of coal spirals
More recently, Bethell and Moorhead (2003) reported plant operating data
showing that combined WOC and spiral circuits offered advantages in terms
of improved separation performance, higher capacity, and lower costs per ton.
Figure 8 provides a simplifed schematic of a typical two-stage WOC and
spiral processing circuit. In this circuit configuration, 1 × 0.15 mm feed coal
is pumped to a bank of WOCs to reject high-density particles. The upgraded
overflow is then directed to a bank of classifying cyclones to deslime and
remove most of the –0.15 mm (100 mesh) material. In this case, the WOCs
must be configured to provide a low-density cutpoint, which typically requires
that they be equipped with a relatively short vortex finder. Lower operating
pressures, e.g., 69–83 kPa (10–12 psi), may also be utilized to ensure that
a low cutpoint density is maintained in the WOCs. The coal-rich particles
intentionally lost to the WOC underflow are then recovered using a bank of
spirals. Spirals are ideally suited to recover the larger, low-density coal particles
contained in the WOC underflow. However, since the underflow stream also
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Combined Water-Only Cyclone and Spiral Circuits 251
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
252 Beneficiation Technologies
very important to ensure that misplaced rock is removed from the final clean
coal product. The secondary middlings product from the spiral units can be
discarded, sent to the clean coal product, or recycled back to the circuit feed.
Generally, the latter option is recommended to improve the overall sharpness of
the separation. The beneficial impact of and scientific rationale for utilizing the
middlings recycle has been described in detail elsewhere (Kohmuench 2000).
The positions of the spiral splitters should be optimized to ensure that an exces-
sive circulating load of middlings is not created within the two-stage circuitry.
Finally, the clean products from the spiral and classifying cyclone overflow
are passed across fine wire sieves to dewater the clean coal. This final process-
ing step also plays a critical role in efficiently removing any residual high-ash
ultrafine clay that would otherwise contaminate the final coal product (Barbee
and Nottingham 2007). In fact, one variation of this type of circuitry passes
the clean product from the spiral to the classifying cyclone sump to provide
an additional step of desliming before final dewatering on the fine wire sieves.
Although this strategy may have some merit, it is not recommended in cases
of high-grind coals that would be subjected to further size degradation in the
classifying cyclone circuit.
The intrinsic benefits of the two-stage WOC and spiral circuit can be
best understood using the illustration provided in Figure 9. The feed matrix
shown in the top left corner of the diagram represents the different types of
coal (black) and rock (gray) particles fed to the circuit. In the case shown, the
WOCs are configured to direct most of the high-density rock to the underflow
stream, with only the finest particles of high-ash material reporting to over-
flow. These particles are easily removed downstream using a combination of
classifying cyclones (to handle the large volume flow) and fine wire sieves (to
efficiently remove bypassed ultrafines). Because of the differential cutpoints in
the WOCs, a considerable amount of low-ash coal in the coarser size fractions
is misplaced to the WOC underflow. Fortunately, feed streams that have been
hydraulically classified by the WOC unit are ideally suited as feeds for flowing-
film separators such as spirals. Since the feed to the spiral contains a dispropor-
tionally large amount of rock, it is important to utilize two-stage compound
recleaner spirals in this application to miminize the unwanted misplacement of
high-ash particles to the clean coal product. The compound spirals efficiently
reject coarse rock without losing a significant amount of lighter coal particles
that are readily carried into the clean coal product by the body drag forces in
the flowing film. The clean product from the spiral contains the coarse low-ash
particles as well as a considerable amount of fines that are hydraulically carried
with the bulk of the slurry flow that reports to clean coal. These unwanted fines,
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Combined Water-Only Cyclone and Spiral Circuits 253
Water-Only Classifying
Cyclones Cyclones
Spiral Sieve
Separators Screens
which contain a significant amount of high-ash clays, are readily removed using
one or more stages of fine wire sieves. The bottom right matrix shows the final
high-quality cleaned and classified product from the combined WOC and
spiral circuit.
P E R F O R M A N C E C O M PA R I S I O N
To better evaluate the capabilities of a combined WOC and spiral circuit, a
case study was undertaken using partition data collected from an industrial site.
Unfortunatley, a direct side-by-side comparision of the different technologies
was impractical because of natural fluctuations in the plant feed and random
errors in the sample collection and laboratory analysis procedures. Therefore,
the experimental data were subjected to mass balancing to obtain a consistent
set of experimental data and then used to obtain size-by-size partition curves
for the WOC and spiral units. The resultant partition curves were used to
simulate the separation performance for an identical set of feed coal washability
data. The simulations included the following types of circuitry:
• WOC only
• Spiral only (spiral second-stage middlings sent to reject)
• Spiral only (spiral second-stage middlings sent to clean coal)
• Spiral only (spiral second-stage middlings recycled to circuit feed)
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
254 Beneficiation Technologies
G U I D E L I N E S A N D R E C O M M E N D AT I O N S
A close examination of the data from this investigation and other sites has
been used to develop several recommended guidelines for the installation and
operation of combined WOC and spiral circuits. The most important of these
include the following:
• The performance of the combined WOC and spiral circuit is very sen-
sitive to the particle size distribution of the feed stream. The presence
of gross oversize particles above the target topsize of 1 mm has a large
adverse impact on separation performance and should be minimzed
to avoid deterioration of product quality and yield. A slip-stream tell-
tell screen is recommended to assist plant operators in detecting this
unfavorable condition.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Combined Water-Only Cyclone and Spiral Circuits 255
80
Spiral Only
(Reject Midds)
74
72
WOC Only
WOC
Spiral
WOC & Spiral
70
7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0
Clean Coal Ash, %
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
256 Beneficiation Technologies
SUMMARY
Two-stage circuits incorporating WOCs and spirals can offer an attractive
method for treating fine coal if properly designed and operated. From a capital-
cost standpoint, utilizing WOCs, with their higher unit processing capacity,
and recleaning the WOC underflow stream with spirals results in a reduced
plant volume for the fine coal circuitry. In addition, overall slurry processing
volumes are reduced when recycing spiral middlings (typically 35%–40% solids
by weight) versus the overflow from second-stage WOCs (typically 5%–10%
solids by weight). Typically in a two-stage WOC circuit, each secondary-stage
cyclone requires an additional primary-stage unit to process the volume of
overflow slurry being recycled. In contrast, recycling spiral middlings, with
their high solids concentration, can typically be handled by the primary-stage
cyclones without additional units.
From a performance standpoint, utilizing a “classifier” type device together
with a “flowing film” separator provides more consistent SG50 values for the
1 × 0.15 mm size fraction than if either unit was used for both the primary
and secondary stages of cleaning. Since maintaining a consistent SG50 by size
is critical to overall separation efficiency, mating a WOC and spiral in the
same circuit can result in improved performance. In addition, recycling spiral
middlings provides a significant improvement in Ep values for individual size
fractions and should always be utilized. Additional advantages in combustible
recovery can also be realized with the circuit. Because the top size of the fine
coal circuit increases as the raw coal screen panels wear, a two-stage WOC
circuit is extremely vulnerable to excessive combustible losses because the
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Combined Water-Only Cyclone and Spiral Circuits 257
REFERENCES
Abbott, J. 1981. The optimisation of process parameters to maximise the profitability
from a three-component blend. In 1st Australian Coal Preparation Conference.
Edited by A.R. Swanson. Newcastle, Australia: Newey and Beath Printers. pp.
87–105.
Barbee, C.J., and Nottingham, J. 2007. Black Bear prep plant replacement of high
frequency screens with fine wire sieves. In Proceedings, 24th Annual Coal Prepa-
ration Conference and Exhibition, Lexington, KY. New York: Penton Media. pp.
113–122.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
258 Beneficiation Technologies
Bethell, P.J., and Arnold, B.J. 2003. Comparing a two-stage spiral to two-stages of
spirals for fine coal preparation. In Advances in Gravity Concentration. Littleton,
CO: SME. pp. 107–114.
Bethell, P.J., and Moorhead, R.G. 2003. Operating characteristics of water-only
cyclone/spiral circuits cleaning fine coal. In Advances in Gravity Concentration.
Littleton, CO: SME. pp. 93–106.
Bull, W.R., Pillai, K.J., and Spottiswood, D.J. 1987. An analysis of water-only cyclone
capabilities. Preprint No. 87-100. Littleton, CO: SME.
Holland-Batt, A.B. 1994. The effect of feed rate on the performance of coal spirals.
Coal Prep. 14:199–222.
King, R.P., Juckes, A.H., and Stirling, P.A. 1992. A quantitative model for the predic-
tion of fine coal cleaning in a spiral concentrator. Coal Prep. 11:51–66.
Kohmuench, J.N. 2000. Improving efficiencies in water-based separators using math-
ematical analysis tools. Ph.D. dissertation, Virginia Polytechnic Institute and State
University, Blacksburg, VA.
Li, M., Wood, C.J., and Davis, J.J. 1993. A study of coal washing spirals. Coal Prep.
12:117–131.
Luttrell, G.H., Kohmuench, J.N., Stanley, F.L., and Trump, G.D. 1998. Improving spi-
ral performance using circuit analysis. Miner. Metall. Eng. 15(4):16–21.
Luttrell, G.H., Catarious, D.M., Miller, J.D., and Stanley, F.L. 2000. An evaluation of
plantwide control strategies for coal preparation plants. In Control 2000. Edited
by J.A. Herbst. Littleton, CO: SME. pp. 175–184.
Luttrell, G.H., Honaker, R.Q., Bethell, P.J., and Stanley, F.L. 2007. Design of high-
efficiency spiral circuits for coal preparation plants. In Designing the Coal Prepa-
ration Plant of the Future. Edited by B.J. Arnold, M.S. Klima, and P.J. Bethell.
Littleton, CO: SME. pp. 73–87.
Mikhail, M.W., Salama, A.I.A., Parsons, I.S., and Humeniuk, O.E. 1988. Evaluation
and application of spirals and water-only cyclones in cleaning fine coal. Coal Prep.
6:53–78.
Osborne, D.G. 1988. Coal Preparation Technology. Vol. 1. London: Graham and Trot-
man. pp. 347–365.
Schlepp, D.D., and Schmidt, M.P. 1988. When to use water-only cyclones. In Indus-
trial Practice of Fine Coal Processing. Edited by R.R. Klimpel and P.T. Luckie.
Littleton, CO: SME. pp. 81–86.
Subasinghe, G.K., and Kelly, E.G. 1991. Model of a coal washing spiral. Coal Prep.
9:1–11.
Subasinghe, G.K., and Kelly, E.G. 1992. Predicting the cut specific gravity of a coal
washing spiral. Miner. Eng. 5(2):193–203.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Moisture Reduction
and Special Topics 4
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Performance, Operation, and
Maintenance Experience of Coal
Ultrafines Filtration with Modern
High-Speed Disc Filters
Jürgen Hahn
ABSTRACT
After a successful modernization and revamp of standard design disc filters in
several coal handling and preparation plants (CHPPs) in Australia (e.g., Saraji
mine), Billiton Mitsubishi Alliance (BMA) decided to go with modern high-speed
disc filters in their new coking coal preparation plant at Blackwater, Australia.
These two Bokela L4 disc filters have been in operation for 4 years, and further
high-speed disc filters have been ordered in the meantime by BMA and Anglo
for new coking coal applications because of the outstanding performance and easy
operation and maintenance. Layout tests for coking coal applications (–0.25 mm)
in the Republic of South Africa (RSA) show similar results compared to Austra-
lia with specific solids throughput rates of 500–1,000 kg/m²/h and moistures in
the low 20 wt% range. Furthermore, the filtration results of the tailings are very
promising as well with specific solids throughput rates of about 500 kg/m²/h and
moistures around 20 wt%. This would allow preparation plants to run high-speed
disc filters for both coal concentrates and tailings and use spare filters for both pur-
poses. In the future, high-speed disc filters may even be used for thermal coal fines
(–0.125 mm).
D E F I C I E N C I E S O F O L D D I S C F I LT E R C O N S T R U C T I O N S
A P P L I E D F O R C O A L F I LT R AT I O N
Vacuum disc filters are the most compact rotary filter type and are comprised
of a series of vertical discs, usually up to 10 or 12 per filter. The conventional
discs, each being composed of up to 24 radial segments, are arranged on a
horizontal shaft. Each segment of the discs is covered with filter media, usually
261
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
262 Moisture Reduction and Special Topics
exchangeable filter cloth. Disc filters can have more than 240 m2 filter area in a
single machine. Main characteristics of disc filters are
• Low footprint area
• Relatively low capital costs (around Euro 2.500–3.000 per m2)
• Low residual moisture content in filter cake
• Low operating costs
• Low maintenance
• Very effective cake discharge system (up to 100%)
• Variations in feed flow rate acceptable
• High specific solids throughputs
• Can handle fairly large particle size distribution of feed material
• Operational problems if no level control available
• Inadequate discharge without snap-blow system
• Inadequate for cake washing
• Older type filters require more maintenance and attendance
Old Design
Looking to old disc filter constructions of 1960s technology, many remarkable
deficiencies can be observed. These weak points can be categorized and sum-
marized as follows.
Process
• Poor cake pickup—the cake is dropped back into the slurry bath as it
emerges out of the bath.
• Cake moistures were higher than expected.
• Poor discharge of the cake caused 50% or more of the cake to fall
behind the scraper into the slurry bath.
• The vacuum was always lower than required for good performance.
Maintenance
• Cloths were tearing on the scrapers, which drove the change to wire
mesh filter media.
• Temporary platforms were required to exchange bats (segments),
which has generated a culture of cloth changes only during plant
shutdowns and weekends no matter how many cloths were damaged.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Ultrafines Filtration with Disc Filters 263
• The design of bat fixing system was cumbersome, which required tie
rods to keep bats in place.
• Bats were heavy, weighing around 40 kg, making the re-clothing task
difficult.
• High wear of filtrate piping and control head due to holes in the filter
media caused high solids in the filtrate.
• Filtrate pipes with 12 to 14 discs were made in two halves with a welded
joint in the middle. This joint breaks within 2 months because of fatigue.
This joint loses vacuum and introduces solids into the filtrate.
• Agitator maintenance is always problematic and often replaced by
compressed air.
I M P R O V E M E N T O F D I S C F I LT E R S B Y S Y S T E M AT I C
REDESIGN
Based on the consequent transfer of fundamental knowledge from the filter
theory and the experience gained in more than 250 filter optimization/revamp-
ing projects for all major filter types and manufacturers, Bokela developed the
Boozer disc filter, a new generation of disc filters incorporating many innova-
tive changes to conventional design practice. Most of these innovations have
been made to resolve capacity and performance problems related to hardware
bottlenecks and/or poor hardware design. The most critical filter parts that
were addressed in the new design are
• Filter discs and segments,
• Filtrate pipes,
• Center barrel and bearings,
• Filter trough,
• Control head, and
• Cake discharge.
The second development focus has been to minimize maintenance require-
ments and time, which has lead to the following results:
• Low wear segments with massive hydraulic capacity to process the
large filtrate flow
• Lightweight segments with quick release bayonet connections for ease
of installation and removal—no tie rods!
• Quick fit filter bag system to allow cable ties to seal the neck
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
264 Moisture Reduction and Special Topics
• Filter designed to use poly bags, which are easier to replace and more
cost-effective
• Permanent walkways to allow easy access to replace segments with
holed cloths
• Removable filtrate pipes
• Center barrel and bearings designed for the high loads encountered
with high capacity
• Gearbox and motor designed for high loads at low speeds
• Filter trough designed to eliminate agitators by being self-agitating
• Level control system to prevent overflow back to feed tank
• Control head with low pressure losses (low wear) at high capacities
• Back suck on the filter cloth to prevent damage on the scrapers during
cake discharge.
Consequently, all the previous changes have resulted in
• Drastically improved cake pickup due to the high vacuum achieved
inside the disc,
• Better cake moistures than other filters of the same area at the same
tonnage,
• Excellent discharge of the cake with 95% to 100% reporting to the
product, and
• Vacuum always at the level required for good performance.
F E AT U R E S O F N E W VA C U U M D I S C F I LT E R
G E N E R AT I O N A P P L I E D F O R C O A L F I LT R AT I O N
With the high-performance disc filter Boozer, Bokela Company has developed
a new generation of large-diameter disc filters, which have set a new standard
for seed filtration in the alumina industry and in the dewatering of coal slurries.
BMA decided to go with these modern high-speed disc filters in their new coking
coal preparation plant at Blackwater, Australia. Two Bokela L4 disc filters have
been in operation for 4 years, and additional high-speed disc filters have been
ordered in the meantime by BMA and Anglo for new coking coal applications
because of the outstanding performance and easy operation and maintenance.
The outstanding hydraulic characteristics of the Boozer disc filter were
achieved by improving each detail of the filter design leading to extraordinarily
high performance capacity, high operational safety and reliability, as well as
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Ultrafines Filtration with Disc Filters 265
1. Filter Segments
2. Center Barrel
3. Control Head 1
4. Filtrate Receiver 2
5. Blowback Tank 3
4
6 5
6. Snap-Blow Valve
7 7. Filter Trough
8. Drive Unit
low maintenance and operation costs. The main features of this new disc filter
generation are listed as follows:
• Minimized pressure drop leading up to 100% higher pressure differ-
ence at the filter cloth compared to conventional disc filters
• Double capacity compared to conventional disc filters
• High-filter speed of 6 rpm
• High operational reliability and flexibility
• Easy maintenance
• Fully automatic and safe operation due to superior process philosophy
realized in a programmable logic controller system
Disc diameters range from 1.7 m to 5.6 m. For filtration of large slurry feed
rates such as tailings, a Boozer filter with a large disc diameter of 5.6 m (L-type)
is the appropriate filter size that is available with one to four filter discs.
The Boozer disc filter is designed by optimizing each detail consequently
according to flow requirements. In this way, pressure loss in the whole flow
route from the filter segments up to the filtrate receivers has been minimized. In
the following sections, the main components of the Boozer disc filter (Figure 1)
are described.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
266 Moisture Reduction and Special Topics
segments, ensuring optimal hydraulic conditions for cake formation, fast filtrate
drainage, and complete cake discharge in less than 0.2 seconds without filtrate
blowing back (i.e., without cake re-wetting). On the other side, this number
of low-weight segments is ideal to facilitate re-clothing and maintenance. The
segment bell is made for a fast, easy, and secure filter bag fixing and is designed
in a shape that allows fixing of the filter bags with cable ties. This fast and
secure method simplifies filter cloth fixing and reduces the time needed for
re-clothing.
Filter Arrangement
Between the control head and filtrate receiver, the previously mentioned
two-phase flow of filtrate and air can lead to a high pressure loss, Δp, of up to
10 times higher than a one-phase flow. Hence, the arrangement of the filter and
filtrate receivers influences filter operation and filter performance. Therefore,
the Boozer disc filter and the filtrate receiver are located as close and straight
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Ultrafines Filtration with Disc Filters 267
as possible to keep this flow section very short (see Figure 1). The layout of the
pipes between control head and receiver avoids any bend as far as possible, and
the receiver inlet is carried out tangentially.
C O M M I S S I O N I N G I S S U E S A N D D E S I G N A D A P TAT I O N S
AT B L A C K W AT E R
At the coking coal preparation plant at Blackwater, two Boozer L4-type disc
filters have been in operation for 4 years. During commissioning of these two
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
268 Moisture Reduction and Special Topics
Single Compartments
Figure 3 Joint single trough (all discs run with the same slurry level)
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Ultrafines Filtration with Disc Filters 269
Figure 4 Complete cake discharge (left) and advanced scraper design (right)
Figure 5 Antifriction inlays for discharge chutes made of UHMPE (left), and disc
guide rollers (right)
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
270 Moisture Reduction and Special Topics
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Ultrafines Filtration with Disc Filters 271
Loop 1
Filter Speed Controlled
by Slurry Level
Loop 2
Vacuum Controlled
by Slurry Level
LIC
Figure 6 Filter operation control system of the high-performance Boozer disc filter
with fully automatic operation
If the slurry level in the filter trough drops or increases because of a chang-
ing slurry flow or changing product characteristics, the filter speed is automati-
cally reduced or increased to maintain a constant high slurry level in the trough
(level set point). If the filter runs at minimum permitted speed and the slurry
level in the trough still drops, then the second control circuit becomes active
and the vacuum in the filter cake formation zone is adapted accordingly.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
272 Moisture Reduction and Special Topics
F I LT E R O P E R AT I O N AT B L A C K W AT E R C H P P
The most important data for the two disc filters are listed as follows:
• Filtration area of 176 m2 each
• Four discs per filter
• Disc diameter of 5.6 m
• Thirty light segments (24 kg each) per disc with bayonet segment fix-
ing for fast change-out
• Filter barrel with internal filtrate pipes
• Joint single trough that works without agitator and allows perfect level
control
• Permanent walkways between the discs
• Snap-blow system for cake discharge (snap-blow valve, control system,
and feed tank)
• Cloth washing system as wash bars underneath scraper with solenoid-
operated shut-off valve
• Automatic level control so there is no overflow recirculation required
• Filter speed n = 0.5 to 4 rpm (typical operating range for coal slurry
filtration)
Maintenance
With the realized design adaptations described previously, the filters run with
an availability of 99%. Scheduled maintenance effort is 31 h/a (1.5 days per
year). Maintenance is required for re-clothing (two times/a, i.e., 3 h for re-
clothing of one disc every 1.5 months), exchange of roller guides (4 h/a), and
maintenance of the snap-blow valve for cake discharge (one per year). In 4 years
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Ultrafines Filtration with Disc Filters 273
Figure 7 Solids throughput vs. square root of filter speed for different feed
concentrations—layout performance for Boozer L-type disc filter with Blackwater
ultrafine clean coal slurry
Table 1 Performance of Boozer L-type disc filter at BMA Blackwater CHPP (ultra-
fine clean coal slurry with solids particle size <300 µm)
Feed Specific Solids Filtrate
Concentration, Flocculant, Throughput, Solids,
wt% g/t DS Moisture, % kg/m2h g/L
Layout 35 minimum <15 <25 517 nominal <10
performance 38 nominal 812 maximum
56 maximum
of operation, only one unexpected filter stop was necessary due to holes in the
control head, but no other abrasion has occurred since then.
R E S U LT S W I T H R S A C O K I N G C O A L — F L O TAT I O N C E L L
OVERFLOW
Coking coal from RSA (ash contents around 10 wt%) is showing good filtra-
tion behavior. The pure flotation cell overflow with a typical feed solids content
of 14–18 wt% is already filtering with a specific solids throughput of 400 kg/
m²/h, which relates to 70 t/h on typically used L4 vacuum disc filters.
Furthermore, the RSA coking coal is very responsive to flocculant dosage.
With 10 g/t flocculant dosage, the specific solids throughput doubles, as shown
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
274 Moisture Reduction and Special Topics
1,200
dp = 60 kPa, No Floc Dosage
Specific Solids Throughput mS, kg/m2h
600
400
200
0
0.0 0.5 1.0 1.5 2.0
1
n
min
Figure 8 Solids throughput vs. filter speed for different flocculant dosages (South
African coking coal—flotation overflow)
in Figure 8 (140 t/h per standard L4 disc filter). The solids throughput further
increases when more flocculant is added.
In addition, filtration rates can be further increased with the use of a coal
thickener. If the flotation overflow is thickened from 15 wt% to about 30 wt%,
the specific solids throughput triples, as Figure 9 shows. The specific solids
throughput increases from about 800 kg/m²/h to more than 2,500 kg/m²/h.
This means that one standard L4 disc filter is able to filter up to 450 t/h.
However, high solids throughput is usually at the expense of higher mois-
ture. Figure 10 shows this very well. A typical ratio of dry time to form time of 2
and a flocculant dosage of 10 g/t (typical setting of standard L4 disc filters used
in coal filtration) is considered for the comparison. Under these conditions, the
nonthickened flotation overflow is filtered with a moisture content of about
20 wt%, while the thickened flotation overflow still contains about 25 wt%
moisture, however, with three times the throughput.
R E S U LT S W I T H R S A C O K I N G C O A L C E L L U N D E R F L O W
Similar to the coking coal concentrate, the coking coal tailings from RSA are
showing good filtration behavior. However, tailings always require thickening
in combination with the addition of flocculant to achieve reasonable through-
put figures.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Ultrafines Filtration with Disc Filters 275
3,000
Specific Solids Throughput mS, kg/m2h
dp = 60 kPa
2,000 10 g/t Floc Dosage
1,500
1,000
500
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
1
n
min
Figure 9 Solids throughput vs. filter speed for thickened and nonthickened flotation
overview (South African coking coal)
30
30 wt% Feed Solids
28
15 wt% Feed Solids
26 Operation Range
24
Moisture Content, wt%
22
20
18
16
14 ∆p = 60 kPa
hc = 15 mm
12 10 g/t Floc Dosage
10
0 1 2 3 4 5 6 7
Dry Time/Form Time [–]
In this case, 30 g/t flocculant was added and the flotation cell underflow
was thickened to a typical thickener underflow solids content of 30 wt%. In this
condition, a specific solids throughput of 460 kg/m²/h (see Figure 11) can be
achieved, which relates to about 80 t/h on vacuum disc filters as used for coal
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
276 Moisture Reduction and Special Topics
500
Specific Solids Throughput mS, kg/m2h
400 ∆p = 60 kPa
30 wt% Feed Solids
350
300
250
200
150
100
50
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
1
n
min
Figure 11 Solids throughput versus filter speed for RSA coking coal tailings
(flotation underflow)
28
30 g/t Floc Dosage
26 Operation Range
24 ∆p = 60 kPa
30% Feed Solids
Moisture Content, wt%
22
20
18
16
14
12
10
0 0.5 1 1.5 2 2.5 3 3.5 4
Dry Time/Form Time [–]
Figure 12 Moisture content versus dry time for RSA coking coal tailings (flotation
underflow)
concentrates. Further thickening and more flocculant addition can increase the
capacity by another 30%–60%.
Within the typical operating range of L4 vacuum disc filters of 1.3 to 4.0
(cake dry time/cake form time), the moisture content of coking coal tailings
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Coal Ultrafines Filtration with Disc Filters 277
700
Specific Solids Throughput mS, kg/m2h
500 ∆p = 60 kPa
Feed Solids = 30 wt%
400 Floc Dosage: 80–120 g/t
300
200
100
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
1
n
min
Figure 13 Test results from different RSA thermal coals—filtration rates
R E S U LT S W I T H R S A T H E R M A L C O A L C O N C E N T R AT E
Vacuum disc filters can also be used for the filtration of thermal coal from RSA.
But thermal coal requires thickening and addition of flocculant to generate a
feed flow suitable to be filtered on vacuum disc filters.
Figure 13 shows two test results from different RSA thermal coals. In both
cases, the coal was thickened to about 30 wt% (typical coal thickener underflow
concentration), and flocculant in the amount of 80–120 g/t was added. This
preparation of the feed flow then allows filtration rates of 300–600 kg/m²/h.
This is a solids throughput of 50–100 t/h on standard L4 disc filters in coal
preparation plants. The moisture depends on the thermal coal properties and
can range from low 20-wt% values to high 30-wt% values, as the test results
show.
REFERENCE
Bott, R., Langeloh, T., and Hahn, J. 2004. Latest state of the art of Al-hydrate seed
filtration. ICSOBA, St. Petersburg, Russia, June 15–18.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Dewatering Fine Coal and
Tailings with a Filter Press
G. Prat
ABSTRACT
The coal industry is facing new challenges with the treatment of ultrafine
(0.044 mm × 0) coal slurries. The need to comply with today’s demanding envi-
ronmental rules, the interest of the industry to recover all valuable products, the
lower moisture requirements, and the need to minimize water consumption have
pushed the industry to develop more efficient dewatering technologies capable of
addressing these challenges. This chapter focuses on filter press technology and its
applications in the coal industry. Three typical applications are presented: raw coal
fines or refuse dewatering, froth flotation concentrate dewatering, and screen-bowl
effluent dewatering. Moreover, representative data from test work carried out in
South Africa in a metallurgical coal plant are presented.
Different options are available when designing a filtration plant utilizing
a filter press as the dewatering method; the selection of the right equipment and
auxiliaries will guarantee a quality product, produced at the lowest moisture and
cost. An overview of the different types of filter presses and auxiliaries is detailed.
INTRODUCTION
Filtration has been and is today a common dewatering method employed for
coal slurries with ultrafine solids. In the past, different types of dewatering
technologies were employed, with rotary drums and disc filters being the most
popular. Today, two technologies are widely established for ultrafine coal dewa-
tering: the filter press and the belt press.
The main application of the belt press in the coal industry is the dewatering
of coal tailings, and this technology is widely used around the world. The belt
press has proven to be a technology capable of treating large volumes of coal
tailings slurry, but on the other hand, it is not able to consistently produce a low
moisture product. In addition, the need for a large amount of polymer dosage
279
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
280 Moisture Reduction and Special Topics
represents a high operating cost per ton of dry coal produced. The demand by
the coal industry for lower-moisture products, reduced treatment cost, and
fewer environmental impacts is putting the filter press in the forefront when
designing a coal tailings dewatering plant. Filter presses have proven to be the
most effective and reliable technique to meet today’s industry requirements.
Filter presses are, however, more expensive in terms of capital costs than belt
presses.
Another field of application for the filter press in a coal preparation plant
is recovery of the filtrate, which exits a screen-bowl centrifuge. This ultrafine
product, sometimes below 10 μm, is especially valuable in metallurgical coal
plants because of its high price. The combination of a screen-bowl centrifuge
and filter press is an ideal solution for total capture of clean coal and to close a
plant’s water circuit (de Korte 2008).
To have a cost-effective, reliable, and trouble-free dewatering plant, it is
critical to make the right selection of filter press and auxiliaries. Various options
are available in the market; an overview of these is detailed in the following
sections.
A P P L I C AT I O N S F O R F I LT E R P R E S S E S AT A C O A L
P R E PA R AT I O N P L A N T
Dewatering 0.15 mm × 0 Raw Coal or Refuse
Dewatering of 0.015 mm × 0 raw coal or refuse (tailings) has been a typical
installation of filter presses in the coal industry around the world (see Figure 1).
This circuit allows the fine raw coal to be recovered to the clean coal product
or for the fine raw coal or refuse to be discarded and disposed along with the
coarse coal refuse. The final moisture of the product ranges from 18% to 30%
(surface moisture) and is largely dependent on size distribution of the feed; the
finer the product, the more moisture retained.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Dewatering Fine Coal and Tailings with a filter press 281
Clear
Water
20%–30%
Solids
Cake
18%–30%
Thickener Buffer Slurry Moisture
Tank Pump
0.15 mm × 0
Froth Flotation
Filter Press
Concentrate
0.15 mm × 0
>15% Solids
Refuse Slurry
Clear
Water
20%–30%
Solids
Clean Coal
Product
Thickener To Impoundment Buffer Slurry 15%–20%
Tank Pump Moisture
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
282 Moisture Reduction and Special Topics
1.5 mm × 0
Clean Coal
Slurry Clean Coal
Screen-Bowl
Centrifuge
Filter Press
Clean Coal
Product
Clear
Water
20%–25%
Solids
Clean Coal
Product
Thickener Buffer 15%–20%
Slurry
Tank Moisture
Pump
Figure 4 shows the results from extensive test work performed on metal-
lurgical coal in a South African coal preparation plant. The test was carried
out with different products, both on concentrate of different size distributions
and on tailings from flotation and run-of-mine refuse. The graph in Figure 4
shows the relationship between “filtration rate” (kg/m2/h) versus the “surface
moisture” (wt%). It is shown that for all applications, the filtration rate was
reduced if the surface moisture was lower. To achieve low moisture, longer air
drying time is needed (see Figure 5), increasing total cycle time and decreasing
productivity. The graph in Figure 5 shows that low moisture can be achieved,
both on concentrate and refuse applications.
Figure 5 shows the influence of air drying time on the surface moisture
level. It can be seen that for longer air drying times, lower surface moistures
were achieved. It also shows the influence of particle size distribution on mois-
ture level. At a higher percentage of –37 µm material, higher moisture contents
were produced for the same drying time. This confirms that size distribution of
the product has a great influence on final product moisture.
T Y P I C A L F I LT E R P R E S S D E W AT E R I N G C I R C U I T
Figure 6 illustrates the typical equipment used in a filter press dewatering cir-
cuit, summarized as follows:
1. The filter press is the main piece of equipment of a dewatering circuit.
2. The slurry pump provides the pressure that enables dewatering. The
selection of the slurry pump, its volume, and type are key to the dewa-
tering process. The slurry pump is responsible for the slurry transfer
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Dewatering Fine Coal and Tailings with a filter press 283
32
Flotation Conc. Ultrafines
30
Flotation Conc. Normal Fines
28 Flotation Tails
ROM T/U
Surface Moisture, % w/w
26
24
22
20
18
16
14
0 20 40 60 80 100 120 140 160 180 200 220 240 260 280
Filtration Rate, kg/m2h
Figure 4 Surface moisture vs. filtration rate for a South African metallurgical coal
81% - 37µ
26
96% - 150µ
86% - 74µ
24
72% - 37µ
22 90% - 150µ
78% - 74µ
20 63% - 37µ
79% - 150µ
64% - 74µ
18 50% - 37µ
16
14
0 40 80 120 160 200 240 280 320 360 400 440 480 520
Drying Time, seconds
Figure 5 Surface moisture vs. drying time for a South African metallurgical coal
into the filter press chambers and for providing the necessary pressure
that produces the solid/liquid separation.
3. The polymer unit is used for preparation and dosing of the flocculant,
which may be needed in some cases when dewatering coal tailings.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
284 Moisture Reduction and Special Topics
Filter Press
Filtrate
Cake
Polymer
Unit
4. The capacity of the buffer tank will depend on the required storage
time and the available space in the plant.
5. The compressor and receiver are used to provide the air for additional
cake drying. Sizing of the compressed air system is especially impor-
tant when low cake moistures are required.
6. The cloth wash system is used for washing the filter cloth.
7. A collecting sump and pump is used to evacuate the potential slurry
that falls to the ground floor, keeping the floor clean and safe.
Filter Press
The filter press is employed for solid/liquid separation using the principle of
pressure provided by a slurry pump. Many options are available in the market,
and it is important to choose the right equipment focusing on
• Simplicity of design and reduced number of moving parts,
• Strong construction,
• Easy operation and maintenance,
• Optional features such as cloth wash and cake available compression,
• Proven design in the coal industry, and
• References.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Dewatering Fine Coal and Tailings with a filter press 285
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
286 Moisture Reduction and Special Topics
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Dewatering Fine Coal and Tailings with a filter press 287
• Light weight
• Welded or detachable membranes
• Can use either air or water as the inflation medium
• Require stay bosses to avoid plate breakage due to differential pressure
• Cloth wear in the stay boss area
• Require cloth washing to remove solids in the sealing area
• Have a limited life, so they need to be replaced due to wear or breakage
• Many models and brands available in the market
Variable plate. Figure 11 shows a variable filter plate. The following list pro-
vides some of the main characteristics for these plates.
• Material of construction: steel core, polypropylene drain plate, and
perimeter rubber seals
• Heavy-duty construction
• Unlimited life
• Watertight chambers
• Flat surface, so no stay bosses required
• Resistant to maximum differential pressure, therefore possible to iso-
late and work with empty chambers
• No cloth required except when dealing with very sticky products
• Replaceable rubber seals
• Variable chamber volume
• Cake compression possible
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
288 Moisture Reduction and Special Topics
Slurry Pumps
Filter press manufacturers have different alternatives when selecting the slurry
pump that feeds the filter press. The most common pumps are given in the fol-
lowing sections.
Centrifugal Pumps
Figure 12 shows two centrifugal pumps used for feeding a filter press. Centrifu-
gal pumps are a common method of slurry pumping in a filter press application.
These units offer a lower cost compared to other pumping methods, which is
the main driving force for manufacturers to select it. The following list provides
some of the main characteristics of a centrifugal pump:
• Electric drive
• Lower cost compared to positive displacement pumps
• Different configurations and linings available
• Wide range of brands available
• Need of variable-speed drive to provide variable flows and pressure
• Alters chemical treatment of the slurry due to floc shear
• Lower filtration pressure compared to positive displacement pumps
• Low energy efficiency
• High wear due to principle of operation
• High operating expense in terms of energy consumption and wear
parts
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Dewatering Fine Coal and Tailings with a filter press 289
Diaphragm Pumps
Diaphragm pumps are an effective method of slurry pumping and as a pres-
sure provider in a filter press application. Compared to centrifugal pumps,
diaphragm pumps provide higher filtration pressures (up to 16 bar), are less
sensitive to cavitation, provide gentle pumping of materials sensitive to shear,
and are more energy efficient. Figures 13 and 14 show two types of piston dia-
phragm pumps.
Piston diaphragm pump. Some of the main features of these pumps are as
follows:
• Electric drive
• Require multiple pistons for high volumes
• Require a variable-speed drive to provide variable flow rates
Hydraulic piston diaphragm pump. Some of the main features of these types
of pumps are as follows:
• Hydraulic drive by the same hydraulic power unit of the filter press, so
low installed power
• Low operating expense; only one moving part
• Variable flow and pressure, so no need for a variable-speed drive
• Operate with upward or downward suction
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
290 Moisture Reduction and Special Topics
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Dewatering Fine Coal and Tailings with a filter press 291
CONCLUSION
The use of filter presses is a proven method for clean coal and tailings dewater-
ing, achieving low moisture consistently with reduced operational costs. To
obtain low operational costs and consistent performance, it is important to
select the right type of filter press and auxiliaries (Bickert 2012). On coal tailings
applications, there might be the need to flocculate the slurry to enhance/enable
filterability. If this is the case, a positive displacement pump is compulsory as
the filter press feeding mechanism to maintain the floc morphology and size.
To obtain low moisture in the final product, drying by means of compressed air
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
292 Moisture Reduction and Special Topics
is required. Filter presses can recover virtually all the solids in the feed, which
make them especially suited to closing the water circuits in plants.
REFERENCES
Alam, N., Ozdemir, O., Hampton, M.A., and Nguyen, A.V. 2010. Dewatering of Coal
Plant Tailings: Flocculation Followed by Filtration. School of Chemical Engineer-
ing, The University of Queensland, Brisbane, QLD, Australia.
Bickert, G. 2012. Filter Presses for Coal and Tailings—Design Details Make Them
Work. ACPS Symposium, Queensland, Australia.
de Korte, G.J. 2008. Dewatering of Ultra-fine Coal with Filter Presses. Report number
CSIR/NRE/MIN/ER/2008/0062/A. Coaltech.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Deep Cone Thickener at Lone
Mountain Processing Plant
ABSTRACT
The Lone Mountain coal processing plant has successfully operated a deep cone
thickener (DCT) since 2007 to concentrate underflow from conventional high-rate
thickeners with 20%–25% solids to 45%–50% solids (by weight). The concen-
trated paste is discharged through a gravity pipeline into the impoundment. Ini-
tially, Lone Mountain discharged the underflow from the conventional thickeners
with 20%–25% solids directly into the impoundment, and subsequently used four
belt presses to concentrate the slurry to 45%–50% solids and trucked the paste to the
impoundment. Replacement of belt presses by the DCT has resulted in savings of
chemical cost, as well as improvement in the plant’s feed rate and operating hours.
The DCT has performed satisfactorily with varying feed density, volumes, and size
distribution of the slurry.
BACKGROUND
The Lone Mountain coal processing plant located in Lee County, Virginia, has
a capacity of 1,200 stph (1,088 mtph). It uses a dense-medium vessel, dense-
medium cyclones, spirals, and column flotation to wash the coal. The coal is
mined from three seams, named Kellioka, Darby, and Owl in Kentucky and
belted to the plant through a tunnel in the Little Black Mountain bordering
Virginia and Kentucky. The slurry from the underflow of two 50-ft (15.2-m)
conventional high-rate thickeners was initially pumped into the impound-
ment. In 1998, slurry from the impoundment leaked through old, unmapped
underground works. To reduce the risk of such an occurrence, Lone Mountain
started concentrating the thickener underflow with 20%–25% solids to 45%–
50% solids (by weight) by belt presses before discharging it into the impound-
ment. The paste from the belt presses was hauled and dumped by truck into
the impoundment. The refuse from the dense-medium vessel, dense-medium
293
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
294 Moisture Reduction and Special Topics
cyclones, and spirals was deposited on top of the paste by trucks and compacted
in layers. Because of high chemical consumption, high maintenance costs, and
frequent bottlenecks associated with belt presses, alternatives to the belt press
were evaluated.
One alternative was the deep cone thickener (DCT), which has been
used successfully to make paste in diamond, gold, and iron ore tailings since
the 1970s but had not been used in coal tailings. After an extensive literature
review, visits to diamond mines in South Africa, and lab tests, a pilot test was
conducted at the Lone Mountain plant. It indicated that a DCT could provide
45% to 55% solids paste and very good quality clarified water for reusing in the
plant’s wash circuit, with a small fraction of chemicals compared to the belt
press operation. After a successful pilot test, the decision was made to construct
a DCT to replace the belt presses.
D E E P C O N E T H I C K E N E R L AY O U T
The Lone Mountain plant, which is located at an elevation of 1,950 ft (594 m),
is approximately ½ mile (0.81 km) from the impoundment. A DCT measur-
ing 50 ft (15.2 m) in diameter and 60 ft (18.3 m) in height was constructed
at an elevation of 2,335 ft (712 m) adjacent to the impoundment by Westek
and was commissioned in mid-May 2007. The underflow from two existing
conventional high-rate thickeners with 20%–25% solids, which was pumped to
the belt presses before, is now pumped to a small sump in the plant at 1,980 ft
(604 m) elevation. It feeds two 6/4 high head Warman centrifugal pumps
working in series. They pump the slurry to the top of the DCT to an elevation
of 2,415 ft (736 m), through a pipeline consisting of 10 in. (254 mm) and 8 in.
(203 mm) diameter polyethylene pipes. Liquid anionic flocculant and liquid
cationic flocculant are added in the feed flume and the center well of the DCT.
Four 7.5 hp (5.6 kW) motors drive a slow moving rake in the DCT. The solids
in the slurry are essentially concentrated by the weight of the slurry. As the rake
rotates, water released from the slurry rises through the space behind the pick-
ets, which support the rake. Clarified water created in the DCT comes back to
the plant by gravity for reuse.
The paste is discharged into the impoundment by gravity through a 10-in.-
(254-mm-) diameter polyethylene pipeline laid at a 14-degree decline. The pool
elevation in the impoundment is approximately 2,170 ft (661 m) at present. It
will rise over time up to 2,290 ft (698 m). A pond was dug approximately 130
ft (40 m) above the DCT and receives water pumped from the underground
mines across the hill in Kentucky through a pipeline in the 2-mile- (3.2-km-)
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Deep Cone Thickener at Lone Mountain Processing Plant 295
long conveyor tunnel. This water is used for the mixing, purging, and makeup
at the DCT and the conventional thickeners at the plant. Figure 1 gives a sim-
plified circuit of the DCT layout. Figures 2 and 3 show the DCT and the slurry
and clarified lines, respectively.
The process at the DCT is operated and remotely controlled from the
plant operator’s control room through a programmable logic controller (PLC)
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
296 Moisture Reduction and Special Topics
system. No one is located at the DCT location. Several cameras also help the
operator to monitor the DCT operation. The slurry feed line has a nuclear
density gauge and flow meter to continually provide the slurry density and
flow rate for the required dosage of liquid flocculants. The PLC calculates and
controls the chemical dosages required. A pressure gauge at the bottom of the
DCT is utilized for the operator to set the pressure limits between which the
DCT needs to operate. The gravity line for paste disposal to the impoundment
also has a flow meter and nuclear density gauge to measure the flow rate and
density of the paste being discharged from the DCT to the impoundment.
Figure 4 shows a simplified circuit of the chemical feed system at the DCT. Fig-
ures 5 and 6 show the gravity discharge pipeline and liquid flocculant pumps,
respectively.
D C T O P E R AT I O N
The underflow streams from two 50-ft- (15.2-m-) diameter conventional high-
rate thickeners are pumped by two centrifugal pumps with variable frequency
drive into a small sump in the plant building. The particle size distribution in
the underflow tailings averages 5% each for 1 mm × 0.25 mm (60 mesh) and
0.25 mm × 0.15 mm (100 mesh), 12% for 0.15 mm × 0.045 mm (325 mesh),
and 78% for 325 mesh × 0. Two high-head centrifugal pumps, working in
series, pump the slurry into the flume of the DCT. The flow rate and density
of the feed slurry are measured continually by a flow meter and nuclear density
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Deep Cone Thickener at Lone Mountain Processing Plant 297
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
298 Moisture Reduction and Special Topics
gauge inputted into the PLC system. The slurry flow rate averages 1,300 gpm
(295 m3/h), and its specific gravity averages 1.13, which equates to 23% solids
(by weight) and 84.5 stph (dry) (76.6 mtph). Anionic and cationic chemicals
for use in the DCT are stored in 5,500-gal (20.8-m3) tanks located beside the
DCT. A 0–3 gpm (0–0.68 m3/h) transfer pump draws anionic chemical from
the storage tank and pumps it to a 3,500-gal (13.2-m3) mix tank after diluting it
with water flowing through a flow meter and dole valve at 80 gpm (18.2 m3/h)
and 85 psi (586 kPa). A 0–20 gpm (0–4.5 m3/h) flocculant feed pump draws
the diluted anionic chemical from the mix tank and pumps it to the top of the
DCT into the flume and the center well. As mentioned previously, the feed
rates of both the anionic and cationic chemicals are varied by the PLC accord-
ing to the incoming slurry’s flow rate and density. Several adjustable windows in
the center well allow the clarified water to flow into the center well and dilute
the incoming slurry to 8%. It was determined that the anionic chemical is most
effective at 8% solids concentration in the slurry.
The solids in the slurry gravitate to the bottom of the DCT where a slow-
moving rake concentrates it to about 50% solids. Water released from this
concentration flows upward through the space created in the paste behind the
vertical pickets by which the rake is suspended. The cationic chemical fed into
the center well clarifies the water sufficiently for reuse in the plant. When the
bed level for solids set by the operator is reached, the PLC automatically opens
the pneumatically operated automatic knife valve in the gravity discharge pipe-
line. A flow meter and nuclear density gauge measure the flow rate and density
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Deep Cone Thickener at Lone Mountain Processing Plant 299
of the paste discharged, and it is recorded continually in the PLC system. The
paste with 45%–50% solids reaches the pool area of the impoundment through
a 10-in. (254-mm) polyethylene pipeline laid at a 14-degree slope. The friction
in the discharge pipeline reduces the shear stress of the paste, and by the time
it comes out of the pipe, it flows like water and has almost zero degree angle
of repose. In the pool area, the DCT paste further consolidates and dries up
by capillary action and evaporation, unlike the paste made by the belt presses,
which forever stayed in paste form because the large amount of chemicals in
the belt press operation inhibited capillary action and thus prevented drying
and consolidation.
After a few months for fine-tuning and tweaking, the DCT has since
worked well and requires very little attention and maintenance. All DCT
operation is PLC controlled remotely without any operational labor. Introduc-
tion of the DCT in the plant circuit has resulted in improvement of the overall
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
300 Moisture Reduction and Special Topics
$0.30
$0.25
$ per Plant Feed, raw tons
$0.20
$0.15
$0.10
$0.05
$0.00
Jul-05
Oct-05
Jan-06
Apr-06
Jul-06
Oct-06
Jan-07
Apr-07
Jul-07
Oct-07
Jan-08
Apr-08
Jul-08
Oct-08
Jan-09
Apr-09
Jul-09
Oct-09
Jan-10
Apr-10
Jul-10
Oct-10
Jan-11
Apr-11
Jul-11
Oct-11
Jan-12
Apr-12
Month-Year
Figure 8 Monthly chemical cost at the Lone Mountain plant for the two conventional
thickeners and deep cone thickener
operating time of the plant compared to belt presses, which had frequent down-
time and caused frequent reduction in coal feed rate to the plant. The capillary
action and consequent drying of the DCT paste in the pool area has resulted
in additional increase in % solids and consequently a lesser volume of the paste.
A paste of up to 1.38 specific gravity equivalent to 55% solids has been pro-
duced by the DCT. However, it is normally operated to produce paste of 1.31
specific gravity (48% solids) to keep the gravity discharge operation trouble free
and to minimize chemical consumption. If there is shortage of makeup water,
the chemical dosage can be readily increased to produce a higher % solids paste.
Figure 7 shows the mass balance of the DCT.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Deep Cone Thickener at Lone Mountain Processing Plant 301
The chemical cost for the plant for both the conventional thickeners and
DCT has averaged about $0.087 per raw short ton fed to the plant compared to
$0.25 for the belt press, in spite of multiple price adjustments in the unit cost of
the chemicals. Figure 8 shows the cost of chemicals at the Lone Mountain plant
in dollars per raw short ton of plant feed for the conventional thickeners with
belt presses until May 2007 and for the conventional thickeners and DCT since
then. The yearly savings in chemical cost has been more than a million dollars.
Additional savings has been realized by discontinuation of four belt presses and
the truck, which was used to haul paste to the impoundment. Overall savings
with the DCT, which is estimated at $1.6 million per year, has resulted in a very
attractive rate of return on the capital invested for the DCT system.
CONCLUSIONS
A DCT to replace belt presses at the Lone Mountain coal processing plant
has performed satisfactorily since mid-2007. The unit is operated to produce
45%–50% solids paste from 100 mesh × 0 tailings, though up to 55% solids
paste has been obtained. Reduction in chemical consumption and maintenance
cost has resulted in a payback of less than 2 years. The paste from the DCT
further increases its % solids level due to capillary action and evaporation in
the impoundment. The paste in the pool area, which was dug by an excava-
tor after the DCT had operated for some time, showed that the top 3–4 ft
(0.91–1.22 m) had solidified to more than 80% solids, but the material below
it from the belt press was still soupy.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Centribaric Operations Update
ABSTRACT
Decanter Machine began developing novel centrifuge technology in January 2008
with the manufacture of a prototype unit that could be used for field study of the
technology. The prototype unit was taken to Walter Energy in April 2008, where
field tests were encouraging at both the No. 7 and No. 4 preparation plants. The first
commercial unit of the technology was purchased by Walter Energy for the Mine
No. 7 preparation plant in September 2009. The unit was started up in February
2010. A single unit was selected to confirm the pilot testing and it required mini-
mal capital to install due to the existing plant layout. The successful installation of
the first unit led to subsequent purchase orders to process the entire effluent streams
for both preparation plants. This chapter discusses the process improvements result-
ing from the installation of the Centribaric units at both preparation plants.
INTRODUCTION
Walter Energy operates the Jim Walter Resources ( JWR) No. 7 mine, a deep
shaft coal mine near Brookwood, Alabama, that produces high-quality metal-
lurgical coal from the Blue Creek seam. Typical raw coal qualities are 90 to
95 HGI (Hardgrove grindability index), and a resulting plant feed particle
size distribution with 30% –1 mm. Raw coal from the mine is hoisted to the
surface where it is stockpiled before being conveyed to the preparation plant.
The original JWR No. 7 preparation plant was designed and constructed by
McNally in the late 1970s and consisted of four parallel circuits processing raw
coal with dense-medium cyclones and conventional froth flotation. Feed to
the dense-medium cyclones was deslimed at 0.5 mm. Flotation cells processed
the –0.5 mm raw coal. Clean coal was dewatered using horizontal vibrating
basket centrifuges and vacuum disc filters. Typical total plant moistures were
10% to 12% moisture, and the vacuum filters typically produced moistures
between 20% and 30%. Contractual requirements for total moisture were often
exceeded using the original process flowsheet. In the late 1980s, the plant was
303
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
304 Moisture Reduction and Special Topics
Refuse
Distributor
Belt
Box
Screen Bowl Screen Bowl
Met Met
Product Product
Raw Fine
Middlings Effluent
Coal Met
Sump Sump Thickener
Sump Sump
Slurry
Pond
Figure 1 JWR No. 7 preparation plant flowsheet of fine coal circuit before
installation of the Centribaric screen-bowl centrifuges
expanded by McNally with the addition of two identical circuits, bringing the
plant feed capacity to 1,200 t/h in six circuits.
In 1996, Sedgman designed and constructed a fine coal circuit upgrade
with the addition of spirals and replacement of the vacuum disc filters with
screen-bowl centrifuges. In the current configuration, raw coal is sized at 1 mm
and 0.15 mm (100 mesh). Raw coal coarser than 1 mm is processed by twelve
711-mm (28-in.) dense-medium cyclones and is dewatered with horizontal
vibrating-basket centrifuges. The 1 mm by 100 mesh size fraction is classified
using cyclones and then processed using thirty triple-start single-stage pri-
mary spirals with the middlings reprocessed in three triple-start single-stage
secondary spirals. Raw coal finer than 100 mesh is processed by six banks of
conventional froth flotation cells. Clean coal from the spirals and flotation cells
are combined and dewatered using five screen-bowl centrifuges. The current
plant feed capacity is 1,400 t/h. The total plant moisture after the Sedgman
fine coal upgrade was completed was approximately 10%, and the screen-bowl
centrifuges typically produce between 12% and 14% moisture. Figure 1 depicts
the JWR No. 7 preparation plant flowsheet prior to the installation of the
Centribaric centrifuges.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Centribaric Operations Update 305
40 100
80
30
Tons per Hour
60
Percent
20
40
10
20
0 0
Refuse Decanter Primary Cyclone Spiral Cell
Stream Effluent Refuse Refuse Tails
TPH 30.0 5.2 2.5 1.8
Percent 75.9 13.2 6.3 4.6
Cum % 75.9 89.1 95.4 100.0
Figure 2 Coal tonnage lost to the refuse stream prior to Centribaric installation
C E N T R I B A R I C C E N T R I F U G E I N S TA L L AT I O N
In March 2011, the Centribaric centrifuge upgrade was complete at the No. 7
preparation plant. The flowsheet calls for the collection of the main effluents
and screen effluents from the standard screen-bowl centrifuges, which are
pumped to a distributor for gravity feeding of the Centribaric units. The main
effluent from the Centribaric units reports to the refuse thickener. The screen
drain from the Centribaric units is recycled. Figure 3 shows the modified flow-
sheet that incorporates the Centribaric centrifuges.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
306 Moisture Reduction and Special Topics
Middlings
Raw Coal Cyclone
Fine Clean
Cyclones
Cyclone
Secondary
Spirals
Float Cells
Distributor
Spirals
Box
Centribaric
Screen Bowls
Refuse Distributor
Belt Box
Met Met
Product Product
Raw Fine
Middlings Effluent
Coal Met
Sump Sump Thickener
Sump Sump
Slurry
Pond
The most immediate and noticeable change with installation of the Cen-
tribaric centrifuges was the immediate reduction in the amount of stable froth
on the refuse thickener. Significant economic benefit of the installation was the
reduction in the amount of solids reporting to the refuse through the effluents
of the centrifuge. The effluent percent solids by weight reporting to the thick-
ener have ranged between 0.5% and 1.0% and have an ash value between 30%
and 50%. This is compared to the standard screen-bowl centrifuge effluents of
4% to 6% solids and 15% ash. Figure 4 depicts the coal losses by refuse stream
after installation of the Centribaric units.
The amount of coal lost to the thickener has decreased from 25 to 30 TPH
to less than 5 TPH of dry solids. The removal of the coal from the refuse stream
has resulted in a decrease in the dry solids pumped to the fine slurry disposal
impoundment by 15%. This equates to an annual volume of approximately
1.02 × 105 m3 (3.6 × 106 ft3) of future disposal capacity per year. The refuse
thickener chemical treatment costs have decreased because of the reduced ton-
nage reporting to the thickener. Figure 5 shows the impact of the Centribaric
unit installation. The thickener underflow ash value increased, reflecting the
removal of the fine coal lost in the effluents of the standard screen-bowl units.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Centribaric Operations Update 307
14 100
12
80
10
Tons per Hour
8 60
Percent
6 40
4
20
2
0 0
Refuse Primary Cyclone Decanter Spiral Cell
Stream Refuse Effluent Refuse Tails
TPH 5.2 4.5 2.5 1.8
Percent 37.1 32.1 17.9 12.9
Cum % 37.1 69.3 87.1 100.0
Figure 4 Coal tonnage lost to the refuse stream after Centribaric installation
80
Underflow
Cell Tails
70
60
Ash Value, %
50
40
20
June 1, 2011 July 14, 2011 August 31, 2011
Sample Date
Figure 5 Comparison of ash values of the thickener underflow and flotation tails
before and after installation of Centribaric screen-bowl centrifuges
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
308 Moisture Reduction and Special Topics
CONCLUSION
The installation of the Centribaric units at the Walter Energy preparation
plants provided significant economic return with the capture of the fine coal
lost to the refuse stream. In addition to the increased coal recovery, other
benefits that can be attributed to installation of the Centribaric units include
extending the life of an existing slurry disposal facility with the capture of fine
coal and reduced chemical consumption for the refuse thickener.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Belt Filter Press in Coal Tailings
Dewatering—A Comprehensive
Economic, Design, and Process
Analysis
Matt S. Fenzel
ABSTRACT
Increasing environmental regulations have created new and changing demands
for coal preparation plant operators. Critical among these is the disposal of fine coal
tailings. Belt filter presses have been utilized as a cost-effective option for dewater-
ing fine coal tailings for ultimate disposal for decades. As regulations have become
more stringent regarding disposal options, plant designers, operators, and equip-
ment suppliers have developed comprehensive economic, design, and process evalua-
tion techniques to ensure proper equipment selection and system design. Plant and
tailings disposal design, material characterization, site-specific dewatering options,
process performance capabilities, and capital and life-cycle costs are important
considerations to ensure a successful tailings dewatering system and disposal design.
INTRODUCTION
Global changes in coal tailings (refuse) disposal, requirements, regulations, and
coal washing processes have required the industry to evaluate the numerous
variables impacting dewatering technology to ensure a sustainable, economic
decision in mine waste disposal. As of 2001, there have been at least 713 active
fresh-water and slurry impoundments in the United States that are monitored
by the Mine Safety and Health Administration. This is in addition to an aver-
age of 70–90 million tons of fine coal tailings that are reported per year (NRC
2002). Considerations of rehabilitation costs, difficulties, and time associated
with slurry impoundments have also caused further evaluations (Harriman
2012). These statistics along with other variables contribute to why tailings
309
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
310 Moisture Reduction and Special Topics
treatment remains the most difficult and expensive area of coal production
(Budge et al. 2000).
Today, it has become increasingly important to provide detailed and
comprehensive analysis in prefeasibility and economic evaluations of not
only whole plant designs, but individual equipment and components, so as
to ensure plant optimization and proper handling of tailings. Current trends
in coal preparation practice and development are influenced by three major
factors of legislation: changes in coal mining practice, quality of raw coal, and
pressures on commercial performance (Budge et al. 2000). In the case of belt
filter press technology, it has been a practical option for decades in fine coal tail-
ings dewatering as well as many other industries worldwide and has assisted in
eliminating problems associated with process waste streams. In the recent past,
dewatering technology options such as recessed chamber and membrane filter
presses have been introduced. However, the belt filter press, when compared
to other mechanical dewatering options, is a low capital cost option (Bickert
2004), with similar operating costs of alternative tailings technologies such as
filter presses (Mathewson et al. 2007). As technology suppliers, design firms,
and end users adjust with the changing variables introduced by new mining
practices, fine coal recovery technology, and regulatory issues, it is important to
perform a site-specific economic and process analysis when selecting the most
appropriate fine tailings dewatering and refuse disposal option.
B E LT F I LT E R P R E S S T E C H N O L O G Y
The belt filter press (Figure 1) is a mechanical dewatering device extensively
used for dewatering solids suspended in water. Since the 1980s, belt filter
presses, also referred to as twin belt filters, belt filters, or multi-roll filters, have
been used to dewater ultrafine coal tailings and other thickened mineral slur-
ries. The goal of the technology is to separate liquids from solids while creating
a product, referred to as “cake,” in a stable, easy-to-handle, stackable solid phase.
Key considerations in the selection of belt filter presses can include
• Continuous operation,
• Sufficient dewatering capabilities under variable conditions,
• Relatively low capital and operating costs, and
• High mechanical reliability and equipment availability.
Operating Principles
The belt filter press is considered a mechanical, pressure filtration, dewatering
device. However, prior to the introduction of feed slurry and the mechanical
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Belt Filter Press in Coal Tailings Dewatering 311
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
312 Moisture Reduction and Special Topics
6. Discharge
5. Stage
High-Pressure
Stage
7. Belt Wash
Stage
Filtrate Water Filtrate Sump
Dewatered Cake
4. Wedge: Upper and lower belts come together to converge on the solid
material to initiate pressure dewatering. External wipers wick away
moisture.
5. High pressure: The solids are compressed and sheared between the
belts while traversing a series of decreasing-diameter S-rolls, releasing
additional water.
6. Discharge: The cake is scraped from the belts, releasing suitable mate-
rial for transport by conveyor.
7. Belt wash: The upper and lower belts cycle through a wash-water box
to ensure a clean dewatering filter belt for the next dewatering cycle.
Process Principles
The belt filter press is typically used to process slurries of fine solid particles
with nearly all particle sizes smaller than 600 µm. The belt filter press is particu-
larly effective over other dewatering technologies when slurries are ultrafine, or
have a high degree of colloidal solids, or have solids that do not settle over time
without polymer addition (Shields 1992).
In belt filter press applications, the small fine particles must be flocculated
during pretreatment (flocculation stage). For fine coal tailings, the typical pre-
treatment requires a dual polymer regiment of an anionic flocculant followed
by a cationic coagulant with a degree of mixing occurring in between stages.
The flocculated structure created by the introduction of an anionic flocculant is
introduced to a small dosage of a cationic coagulant that will bind any remain-
ing finely dispersed clays, while breaking down and tightening the existing
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Belt Filter Press in Coal Tailings Dewatering 313
flocculated structure that will aid in releasing the remainder of its interstitial
water when subjected to pressure from the belt filter press (Budge et al. 2000).
The anionic flocculant is typically a medium-weight, medium-charge-density
polyacrylamide chained molecule that is received in either dry (powdered)
form or as an emulsified liquid (emulsion).
It is also necessary for the anionic polymer to be correctly prepared. It must
be educed and made with clean water, then stirred, separated, and aged by a
polymer preparation system to ensure maximum effectiveness and efficiency
of the polymer. This preparation will require time, suitable water quality, and
an efficient polymer system to provide polymer/water blending, mixing, and
aging. Too much shear or too fast activation will result in either poor floccula-
tion or unnecessary cost of operation (Alderman 2010).
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
314 Moisture Reduction and Special Topics
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Belt Filter Press in Coal Tailings Dewatering 315
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
316 Moisture Reduction and Special Topics
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Belt Filter Press in Coal Tailings Dewatering 317
•
Suppliers will sometimes underestimate operating costs.
•
Plant operations may not be optimized.
•
Export data sometimes cannot be directly compared.
•
Coal tailings differ significantly and thus performance and other char-
acteristics will also change depending on the material type (Bickert
2004).
By considering some of the expenditures associated with dewatering equip-
ment that include capital costs and operating expenses such as wear parts, labor,
power consumption, and chemical consumption, a study will be able to deter-
mine a close estimate of how much it would cost to install, operate, and main-
tain a belt filter press at a specific site. In conjunction or even before performing
economic analysis of belt filter press technology, a careful plant and process
design analysis should be performed to accurately size the dewatering system.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
318 Moisture Reduction and Special Topics
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Belt Filter Press in Coal Tailings Dewatering 319
Thickener Polymer
Addition Fine Coal
Tailings
Clarified Water
(<200 ppm)
Polymer
Addition
Dewatered Fines
(Cake Conveyor)
Pump
• 60 dry t/h
• 580 gpm
Established laboratory or pilot demonstration equipment capacity results for
specific material:
• 8 dry t/h/m of belt width
• 77 gpm/m of belt width (assumed density of 35% w/w)
In the assumption that this particular plant would utilize a 3.0-m belt filter
press, the capacity of the machine would be calculated by
t t gal
8.0 h 3.0 m 24
h
230
min
m machine = machine or machine
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
320 Moisture Reduction and Special Topics
t gal
60 h 580 min
= 2.5 machines or = 2.5 machines
24 t 230 gal
h min
machine machine
Rounding both results to the nearest whole machine, this application would
need three 3.0-m belt filter press units to meet peak capacity, without consid-
erations given to system redundancy or maintenance plans. After determining
the quantity of machines, it will also become important to determine the plant
layout to optimize floor space, auxiliary equipment usage, and overall process
for the belt filter press area of the plant. Typical coal preparation plant designs
of tall structures with multiple levels, low portal frame building structures,
or modular plants is a function of available budget, mine life, power cost,
and operating philosophy (Budge et al. 2000). These design philosophies can
influence the selection criteria and decision for what type of filter layout can
be incorporated. Figure 4 shows a cluster of four belt filter presses in line and
discharging onto a common conveyor.
With the installation of a belt filter press or any other mechanical dewa-
tering unit, the number of machines must be sufficient for the plant’s needs
for disposal of reporting waste fines. However, depending on the economic,
production, and process climate of a plant, it may require management and
operators to look at adjusting processes to achieve the necessary result for
waste disposal, especially where cake moisture is important for compaction
and disposal. One approach has been the blending of coarse and fine rejects.
This is commonly referred to as co-disposal and the advantages are as follows
(Bickert 2004):
• Maximizing storage space
• Improved water return from impoundments
• Facilitation of impoundment rehabilitation
• Reduction in the risk of spontaneous combustion of coarse rejects
Alternative and less common approaches have been used to improve cake
moistures and fine tailings dewatering results. Blending or bleeding a small
portion of coarser product rejects into waste fines streams has been used to
provide improved dewaterability. This is achievable through decreasing the effi-
ciency of cyclones or other separating equipment upstream to create a coarser
feed reporting to the waste fines circuit. In general, lower cake moistures are
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Belt Filter Press in Coal Tailings Dewatering 321
typically obtained with more hydrophobic feeds, and increased levels of coarse
material lead to lower cake moistures and higher capacities (Mathewson et al.
2007). However, a plant must again take overall economic considerations of the
small percentage loss in production to the operating costs associated with dif-
ficult processing conditions where this modification would prove to be benefi-
cial. Another alternative that has been used is blending concrete, lime, or waste
by-products (i.e., fly ash) with the discharge cake. These agents typically assist
in the stabilization and handling for coal tailings. Many of these techniques
can be utilized when there is a large quantity of ultrafine, expandable clays in a
plant’s process.
In plant design, whether new or existing, it is also important to accurately
size and/or build redundancy whenever possible in the tailings dewatering sys-
tems. During feasibility studies or evaluations, thorough empirical analysis and
laboratory simulations on the most difficult material to process are necessary to
give proper sizing for the plant to adequately handle this material type. With
correct sizing from laboratory simulations, sound decisions can be formulated
as to whether the plant should install additional machines or if there is an
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
322 Moisture Reduction and Special Topics
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Belt Filter Press in Coal Tailings Dewatering 323
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
324 Moisture Reduction and Special Topics
Pilot Demonstrations
In cases where site specifics generate concerns regarding fine tailing dewater-
ability and compactability for disposal, commercial-scale dewatering equip-
ment pilot studies are recommended (see Figure 6). Pilot programs can vary in
length depending on the number of variables considered. One to two weeks is
typical for generating enough data to establish process and dewatering equip-
ment selection; although in cases where multiple seams are mined, demonstra-
tions can last much longer.
The data shown in Figure 7 are typical of the fines dewatering informa-
tion generated in a commercial-scale pilot study. The overall results indicated
an average capacity range of 5–7 t/h/m to maximum capacity rates of 10–12
t/h/m. Cake moistures were well below 35% w/w moisture, even at maximum
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Belt Filter Press in Coal Tailings Dewatering 325
CONCLUSIONS
Belt filter press technology has been a practical, cost-effective option for
mechanically dewatering coal tailings for decades and is still considered an
excellent solution for a broad range of application conditions. It is important
that equipment suppliers, design firms, and coal companies perform in-depth
evaluations on plant design, especially for the tailings streams. Not only is
this area already one of the most costly operating expenses of a plant, it could
restrict plant production or further increase the cost of handling waste fines if
not properly designed. Both process and economic evaluations should be per-
formed to accurately determine whether a belt filter press is the most suitable
technology for the application or process. As long as designs take a comprehen-
sive and detailed approach in determining equipment selection, coal prepara-
tion plants will be able to adequately and efficiently process coal tailings circuits
and meet disposal requirements.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
326 Moisture Reduction and Special Topics
39.0
Seam 1
37.0
Seam 2
Seam 3
35.0 Seam 4
Seam 5
Total Moisture, %
Seam 6
33.0
31.0
29.0
27.0
25.0
0 2.0 4.0 6.0 8.0 10.0 12.0 14.0
Dry Feed, tph
Figure 7 Capacity and cake moisture results relating to a coal seam in Queensland,
Australia
REFERENCES
Alderman, K. 2009. Belt filter presses. In Coal Prep Primer. Pittsburgh: Coal Prepara-
tion Society of America.
Bickert, G. 2004. Review of Tailings Dewatering Technologies for Australian Coal Pro-
cessing Plants. Final Report to Australian Coal Association Research Program.
ACARP Project C14012.
Bickert, G., Selomulya, C., Liao, J.Y.H., and Amal, R. 2007. Coal fines filtration—The
relevance of coal floc micro-properties. In Proceedings, Eleventh Australian Coal
Preparation Conference. Paper C2. Sunshine Coast, QLD, Australia.
Brown, E.T. 2003. Water for a sustainable minerals industry—A review. In Proceedings,
Water in Mining 2003. Melbourne: The Australasian Institute of Mining and
Metallurgy.
Budge, G., Brough, J., Knight, J., McNamara, L., and Woodruff, D. 2000. Review of the
Worldwide Status of Coal Preparation Technology. Final Report to the Department
of Trade and Industry’s Cleaner Coal Technology Programme. Report No. Coal
R199 DTI/PUB URN 00/1205.
Cousins, B. 2012. Alternatives to coal mine tailings impoundment—Evaluation of
three dewatering methods at Rockspring coal mine. In International Symposium
on Water in Mineral Processing. Englewood, CO: SME.
Harriman, A. 2012. ACPS Dewatering Workshop 2012: Background and Fundamen-
tals. Australian Coal Preparation Society, Newcastle, NSW, Australia.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Belt Filter Press in Coal Tailings Dewatering 327
Mathewson, D.J., Norris, R., and Dunne, M.J. 2007. Tailings dewatering, dry screening
and water clarification for the coal industry. In Proceedings, Eleventh Australian
Coal Preparation Conference. Paper C3. Sunshine Coast, QLD, Australia.
NRC (National Research Council). 2002. Coal Waste Impoundments: Risks, Responses,
and Alternatives. Washington, DC: National Academy Press.
Sanders, G.J. 2007. The Principles of Coal Preparation. 4th ed. Newcastle: Australian
Coal Preparation Society.
Shields, G. 1992, Reducing the Moisture Content of Clean Coals. Vol. 3: Belt Filter Press.
CQ for Electric Energy Research Corporation.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Drying and Plant
Profitability
ABSTRACT
This chapter describes the increased plant revenue that can be achieved by coal
preparation plants by using a suitable drying technology to significantly (by ~50%)
lower the product moisture content of the fine clean coal fraction. The equalization
of incremental inert (i.e., combined ash and moisture) content approach and real
plant feed washability data for an Illinois Basin coal have been utilized to estimate
the potential increase in plant yield that can be achieved for a simple two-circuit
plant by the above-mentioned reduction in moisture content of the fine clean coal
product. Computer simulation results indicate that the overall plant yield of a
typical coal preparation plant can be potentially increased by 5.74% if the mois-
ture content of the clean coal product can be further reduced by 50% (i.e., from
the ~18% level commonly achieved by the best available mechanical dewatering
processes, such as screen-bowl centrifuges, vacuum disc filters, or plate-and-frame
filter presses, to ~9% with the use of a suitable fine coal drying technology). With
the substantial reduction in moisture content and thus the inert content of the fine
clean coal product, the specific-gravity cut of the coarse coal circuit can be increased
from 1.47 to 1.69, resulting in the majority of the above-mentioned yield improve-
ment for the overall plant while still maintaining the desired heat content specifica-
tion of the total clean coal product.
The Parsepco Drying Technology (PDT) and Nano Drying Technology
(NDT), two recently developed drying technologies, have shown the feasibility of
providing the above-mentioned moisture reduction of the fine clean coal product.
The PDT dewaters the moist coal on a woven steel belt fitted with medium-wave
infrared radiators in a negative-pressure environment, whereas the NDT utilizes
molecular sieves to absorb the excess moisture from fine coal. These two emerging
drying technologies along with several other commercially available coal drying
technologies, more commonly used for low-rank coal, have been reviewed in this
329
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
330 Moisture Reduction and Special Topics
chapter for their potential integration into higher rank coal preparation plant
circuits.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Drying and Plant Profitability 331
Drying Rate
Moisture
Drying
Change
Rate Drying Rate
Drying Time
Initial
Period Constant Rate Period Falling Rate Period
Irrespective of the drying method used, each fine coal type will have a
drying characteristic curve as a function of drying temperature, hot air (gas)
velocity, and pressure environment. A typical drying curve (shown in Figure 1),
as described by De Korte and Mangena (2004) indicates three distinct time
periods: the initial increasing drying rate period followed by a constant drying
rate period and a falling drying rate period. As shown, the moisture reduction
rate significantly decreases after the removal of all surface moistures in the con-
stant drying rate period.
During the initial period, the wet coal is heated from the ambient tem-
perature to the process temperature maintained inside the dryer. Heat energy
is transferred to the coal particles, resulting in evaporation of the contained
moisture. The rate of evaporation and thus the drying rate increases rapidly
with the removal of most of the surface moisture during this initial period.
This causes the exponential decay in the moisture content during this period.
At the end of this initial period, when the heat transferred from the source
(hot air, gas, etc.) becomes equal to the the cooling caused by evaporation of
surface water from coal particles, the drying rate stops increasing and continues
at a constant rate throughout this second period. This is shown as the constant
drying rate period with a horizontal line for the drying rate and and a straight,
though inclined line with the same slope over this entire second period for the
moisture content change. When all the the surface moisture from coal particles
is evaporated, drying of the inherent moisture begins, which is marked by a
substantially reduced drying rate in the falling rate period, shown in Figure 1.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
332 Moisture Reduction and Special Topics
The cost of drying the inherent moisture becomes prohibitively expensive due
to the exteremly slow drying rate observed while removing moisture from
the micropores and microcapillaries of individual coal particles. The authors
believe that a suitable drying technology can be used to remove almost all the
surface moisture content of the mechanically dewatered fine coal product in a
coal preparation plant at a relatively low cost. This is how the moisture content
of the fine clean coal can be brought down to almost the same range of values
as that of mechanically dewatered coarse coal. It is true that such change in
commercial practice will require significant capital investment and increase
operating costs for coal preparation plants. However, the resulting reduction in
moisture content of the fine clean coal product would allow suitable increase
in the specific-gravity cut achieved in both coarse and fine coal cleaning circuits
and thus significantly increase the plant clean coal yield while maintaining the
heat content of the plant product at the original level. It is believed that the
resulting increase in plant revenue will far offset the additional cost of integrat-
ing the suitable drying technologies to a conventional coal preparation plant
and, thus, increase plant profitability.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Drying and Plant Profitability 333
preparation plant to nearly the same level as that of the mechanically dewatered
coarse coal product.
To explain the potential benefits that could be realized in a realistic plant
environment, a computer simulation exercise has been conducted using the real
feed washability data obtained from a coal preparation plant operating in the
Illinois Basin. To keep the calculations relatively simple, only two cleaning circuits
(i.e., coarse and fine coal circuits of the plant) have been considered for this analy-
sis. The plant cleans the +2 mm size coarse coal and 2 mm × 75 μm size fine coal
using dense-medium cyclones and coal spirals, respectively. The two-circuit plant
cleans 900 tph (tons per hour) of raw coal having an overall ash value and a mois-
ture content of 21.85% and 5.65%, respectively. The feed washability data for
both coarse and fine coal are listed in Table 1. The as-received (ar) heat content
of the feed coal has been calculated based on the strong correlation between the
heat content (Btu/lb) and the combined ash and moisture content of a variety of
Illinois Basin coals, shown in Figure 2. The product moisture contents of coarse
and fine clean coal for the simulation exercise have been assumed to be 6% and
18%, which are quite comparable to the moisture content that is commercially
achieved in the coal preparation plants in the Illinois Basin.
The computer simulation exercise was targeted to maximize the plant
yield while producing an overall product heat content of 11,900 Btu/lb (on
as-received basis). The equalization of incremental inert content (combined
ash and moisture) approach of plant optimization was pursued for two differ-
ent cases: the first case being the conventional plant without using any drying
technology, whereas the second case made use of a suitable drying technology
to lower the moisture content of mechanically dewatered fine coal by 50% (i.e.,
from 18% to 9%). For both cases, the coarse clean coal moisture content was set
at 6%, which is quite comparable to the moisture content that is commercially
achieved in coal preparation plants in the Illinois Basin.
Figures 3 and 4 illustrate the clean coal yield versus the overall inert con-
tent and incremental inert content relationships obtained for both case 1 and
case 2, respectively. For case 1, the incremental inert content for both coarse
coal and fine coal circuits was equalized at 25.5% to achieve the desired heat
content >11,900 Btu/lb for the overall plant product. As indicated in Table 2,
the heat content of the individual products from the coarse coal and fine coal
circuits were 12,081 Btu/lb and 10,799 Btu/lb, respectively, and the total clean
tonnage produced was 673 tph. The significantly lower heat content of the
fine coal product was the result of the high moisture content of 18% of this
product. Case 2 simulations show the increase in clean coal production that
can be achieved by lowering the fine coal moisture content by 50% (i.e., from
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
334 Moisture Reduction and Special Topics
Table 1 Washability data for the coarse and fine coal obtained from a coal
preparation plant operating in the Illinois Basin
16,000
14,000
12,000
Heat Content, Btu/Lb
6,000
4,000
2,000
0
0 10 20 30 40 50 60 70 80 90 100
% Ash + % Moisture
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Drying and Plant Profitability 335
100
A
90
80
Clean Coal Yield, %
70
60
50 Fine Circuit
Coarse Circuit
40
10 20 30 40 50 60 70 80 90
Incremental Inert Content, %
100
B Fine Circuit
Coarse Circuit
90
80
Clean Coal Yield, %
70
60
50
40
10 12 14 16 18 20 22 24 26 28
Average Inert Content, %
Figure 3 Clean coal yield vs. incremental inert (moisture + ash) and overall inert
content relationships for the coarse and fine coal circuits of a coal preparation plant
not using any coal drying technology
of the overall clean coal product at the original level of 11,937 Btu/lb. Thus,
the additional 52 tph of clean coal can be produced by suitably increasing
the specific-gravity cuts of both coarse and fine coal circuits to 1.69 and 1.74,
respectively. In spite of the significant increase in the incremental inert content
of both circuits to a 39.1% level, the overall inert content of both products and,
thus, the overall heat content could be maintained at the original level. At a
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
336 Moisture Reduction and Special Topics
90
A
80
Clean Coal Yield, %
70
60
50
Fine Circuit
Coarse Circuit
40
10 20 30 40 50 60 70
Incremental Inert Content, %
90
B
80
Clean Coal Yield, %
70
60
50
Fine Circuit
Coarse Circuit
40
0 5 10 15 20
Average Inert Content, %
Figure 4 Modified clean coal yield vs. incremental inert (moisture + ash) and overall
inert content relationships for the coarse and fine coal circuits of a coal preparation
plant simulated with the use of drying technology to further lower the moisture
content of the fine clean coal product from 18% to 9%
rate of $50/t of Illinois Basin clean coal and 6,000 working hours per year, the
additional clean coal production would result in an increase in annual revenue
of $15.6 million. This will require capital investment in a suitable technology to
dry nearly 100 tph mechanically dewatered fine clean coal from 18% moisture
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Drying and Plant Profitability 337
Table 2 Potential increase in clean coal tonnage that can be achieved by integrating
a drying technology into the fine coal circuit of a simple two-circuit plant preparing
Illinois Basin coal
tph tph Moisture, Ash, Btu/lb Inert, Incremental
Coal (dry) (ar) % (ar) % (dry) (ar) % (ar) Inert, % (ar)
Without Drying Technology
Coarse 561.4 597.2 6 8.76 12,081 14.8 25.5
Fine 62.2 75.9 18 5.01 10,799 23.0 25.5
Total clean coal 623.6 673.1 7.4 8.39 11,937 15.7 25.5
With Drying Technology
Coarse 588.6 626.19 6 9.53 11,961 15.5 39.1
Fine 89.7 98.6 9 7.68 11,783 16.7 39.1
Total clean coal 678.3 724.8 6.4 9.29 11,937 15.7 39.1
C O M M E R C I A L LY AVA I L A B L E D R Y I N G T E C H N O L O G I E S
Most of the commercially used coal drying technologies are of convective ther-
mal drying type, in which drying occurs when a hot gas (air) is allowed to be in
contact with moist coal. The circulating hot gas (air) also acts as a carrier for the
removal of evaporated moisture from the dryer (De Korte and Mangena 2004).
Drying of coal is more widely practiced commercially for low-rank coals having
very high inherent and surface moisture contents. Pikon and Mujumdar (2006)
provide a detailed discussion on various commercially used coal dryers in the
Handbook of Industrial Drying. Jangam et al. (2011) provides a comparative
analysis (Table 3) of the advantages and disadvantages of state-of-the art coal
drying technologies. Various past studies (Pikon and Mujumdar 2006; Bongers
et al. 1998; Wilson et al. 1997; Suwono and Hamdani 1991; Mujumdar 1990)
indicate the significant advantages associated with superheated steam drying.
These include reduced risk of fire hazard/spontaneous combustion due to the
absence of oxygen, increased drying rate and energy efficiency, and reduction in
dust emission. These types of superheated steam dryers could be very suitable
for large-scale fine coal drying applications at coal preparation plants.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
338 Moisture Reduction and Special Topics
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Drying and Plant Profitability 339
To Blower
Vacuum/Blower
Dry Coal
Condensate
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
340 Moisture Reduction and Special Topics
Feed
Product
In cases of extremely fine clean coal below 75 µm, a pin mixer (shown in Fig-
ure 6) is used to prepare the feed for the dryer in the form of 3–6 mm size
microgranules. More details about the PDT is available elsewhere (Buisman
2010), which reports some of best drying results achieved from the Parsepco
dryer for extremely fine (–45 µm) coal tailings. Product moisture contents of
9.51% and 13.73% were achieved by drying the dewatered tailings product
obtained from a plate-and-frame filter press.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Drying and Plant Profitability 341
Dry high-moisture
Separate dewatered Reuse molecular
coal fines by
coal fines from sieves after thermal
combining with
molecular sieves. regeneration.
molecular sieves.
Feed Coal/Sieve
Coal Contactor
Water Sieve
Vapor Regenerator
2012). The same study reported product moisture contents in the range of 5%
to 10% for both –0.6 mm and –0.15 mm coal having feed moisture contents in
the range of 22% to 28%.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
342 Moisture Reduction and Special Topics
CONCLUDING REMARKS
The wide-scale practice of drying coal to lower its moisture content has been
restricted mostly to low-rank coal having a significantly higher proportion of
inherent moisture, which cannot be removed by mechanical means. However,
results reported in this chapter, based on a detailed plant optimization analysis
conducted using the approach of equalization of incremental inert content of
each cleaning/dewatering circuit of a two-circuit plant, indicates that suitable
coal drying technologies should also be integrated to the bituminous coal and
anthracite preparation plants of the future. Some of the emerging drying tech-
nologies, based on superheated steam, infrared heating, microwave heating,
and/or molecular-sieve-based nanotechnology drying may be quite useful in
lowering the moisture content of the mechanically dewatered fine clean coal
product by removing only its residual surface moisture content. Attempting to
lower inherent moisture content of fine coal may not be quite viable in most
cases, due to the extremely slow drying rate achieved during the last stage of
drying period (i.e., the falling rate period). However, the residual surface mois-
ture content of fine clean coal could be nearly eliminated by the use of high-
efficiency emerging drying technologies while adding to the profitability of the
coal preparation plant and, thus, mining operations.
REFERENCES
Bland, R.W., Harsh, P., Hurley, M., Jones, A.K., Vinod, K., and Sikka. K. 2011. U.S.
Patent Application Publication, Pub. No. US 2011/0078917 A1 Pub. April 7.
Bongers, G.D., Jackson, W.R., and Woskoboenko, F. 1998. Pressurized steam drying
of Australian low rank coals Part I. Equilibrium moisture contents. Fuel Process.
Technol. 57:41–54.
Bratton, R., Ali, Z., Luttrell, G.H., Bland, R., and McDaniel, B. 2012. Nano drying
technology: A new approach for fine coal dewatering. In 29th Annual Interna-
tional Coal Preparation and Exhibition and Conference Proceedings. Lexington,
KY. pp. 97–110.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Fine Coal Drying and Plant Profitability 343
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
344 Moisture Reduction and Special Topics
Mujumdar, A.S., and Jangam, S.V. 2011. Drying of Low Rank Coal. Minerals, Metals
and Materials Technology Center (M3TC) Report-M3TC/2011/01. National
University of Singapore.
Osman, H., Jangam, S.V., Lease J.D., and Mujumdar, A.S. 2011. Drying of Low-Rank
Coal—A Review of Recent Patents and Innovations. Minerals, Metals and Materials
Technology Center (M3TC) Report-M3TC/TIPR/2011/02. National Univer-
sity of Singapore.
Pikon, J., and Mujumdar, A.S. 2006. Drying of coal. In Handbook of Industrial Drying.
3rd ed. Edited by A.S. Mujumdar. Boca Raton, FL: CRC Press. pp. 993–1016.
Suwono, A., and Hamidani, U. 1991. Upgrading the Indonesia’s low rank coal by super-
heated steam drying with tar coating process and its application for preparation of
CWM. Coal Prep. 21:41–54.
Wang, J., and Sheng, K. 2006. Far-infrared and microwave drying of peach. LWT-Food
Sci. Technol. 39:247–255.
Wilson, W.J., Walsh, D., and Irvin, W. 1997. Overview of low rank coal drying. Coal
Prep. 18:1–15.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Nano Drying Technology—A
New Approach for Fine Coal
Dewatering*
ABSTRACT
The removal of moisture from fine coal has been a long-standing problem in the
coal preparation industry. Although coal fines often represent as little as 10% of
the total run-of-mine feed, this size fraction may contain more than one third
of the total moisture in the final marketed product. Existing thermal dryers can
effectively reduce moisture; however, these massive units require very large capi-
tal expenditures and have become a target of increased environmental scrutiny.
Likewise, existing mechanical equipment for fine coal dewatering tends to produce
unacceptably high moistures that often cannot be tolerated on existing coal con-
tracts. In light of these issues, an innovative mechanical-thermal dewatering pro-
cess known as Nano Drying Technology (NDT) has recently been developed. This
chapter (1) reviews the working features of this novel drying process, (2) presents
experimental results obtained from recent laboratory and pilot-scale test programs,
and (3) discusses the potential advantages of the process over existing thermal dry-
ing and mechanical dewatering systems.
INTRODUCTION
Essentially all coal supply agreements impose strict limitations on the amount
of moisture contained in the shipped product. Residual moisture lowers heat-
ing value, increases transportation costs, and can create downstream handling/
freezing problems for customers. To meet the moisture specification, a variety
of solid–liquid separation processes are used in modern coal preparation plants.
* All rights, title, and interest to this work, including the copyrights, belong to Nano
Drying Technologies, LLC. Nano Drying Technologies, LLC have given permis-
sion to SME, Inc., to include this work in this book.
345
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
346 Moisture Reduction and Special Topics
Available methods for reducing surface moisture can be broadly classified into
three main groups: sedimentation, filtration, and thermal drying (Wills and
Napier-Munn 2006). Sedimentation methods make use of static or induced
centrifugal forces to separate solids from water based on differential settling/
compaction, whereas filtration methods trap particles against a mesh or porous
medium to separate solids from water. Equipment such as vibrating screening
systems and various types of centrifugal dryers (stoker, screen-scroll, and vibra-
tory centrifuges) are commonly used to dewater coarser coal particles. Finer
coal particles (<0.5–1 mm topsize) are typically dewatered using more complex
dewatering equipment such as screen-bowl centrifuges and various types of
vacuum disc and belt filters (Luttrell et al. 2007). Unfortunately, existing fine
coal dewatering processes are inefficient in terms of moisture reduction, solids
recovery, and/or energy consumption (Osborne 1988; Le Roux et al. 2005;
Keles 2010).
It is widely recognized that the moisture content attainable by mechanical
dewatering systems is strongly dependent on coal particle size. For example,
Figure 1 shows the approximate lower limit on moisture than can be attained
using mechanical coal dewatering equipment. The inverse relationship between
particle size and moisture content should be expected because of the sharp
increase in surface area as particle topsize is reduced. The finest coal fraction can
account for as little as a few percent by weight of the total run-of-mine coal but
may represent one third or more of the total moisture in the final coal product.
In some industrial operations, fine (<100–200 µm) or ultrafine (<40–50 µm)
coal particles may be intentionally removed by classification circuits and dis-
carded at the plant site to avoid an unacceptably high product moisture. This
loss represents a waste of valuable coal resources and a potential environmental
liability when discarded into waste impoundments (Orr 2002).
Historically, thermal dryers have been utilized in the coal preparation
industry to reduce clean coal moisture to single-digit values whenever mechani-
cal dewatering processes were incapable of meeting contract specifications.
The most popular design is the fluidized bed dryer, which uses coal, oil, or
natural gas as the fuel source to heat the incoming air stream. The amount of
fuel required depends on the amount of water that must be evaporated, which,
in turn, depends on the amount of coal fed to the dryer and the percentage
of water in the dewatered product (Miller 1998). When operating correctly,
thermal dryers can reduce the clean coal moisture to less than 6% by weight
(Meenan 2005). Unfortunately, thermal dryers involve a substantial investment
of upfront capital funds when installed and large annual costs for equipment
maintenance and repair throughout their life span. Operating costs for thermal
dryers have also greatly increased in recent years in response to higher fuel and
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Nano Drying Technology: A New Approach 347
100
80
60
40
30
Vacuum/Pressure Filters
Product Moisture, %
20
Screen-Bowl Centrifuges
Screen-Scroll
10 Centrifuges (Fine)
8
Vibratory Centrifuges
6
Sreen-Scroll
4 Centrifuges (Coarse)
3 Vibratory Stoker Vibrating
2 Centrifuges Screens
0
325M 200M 100M 48M 28M 14M 8M 6M 4M ¼″ ½″ 1″ 2″ 4″ 8″
Particle Size, inches or mesh
labor costs. Thermal dryers can also suffer from emission problems associated
with fugitive dust and poor opacity. In fact, the opacity standard for coal dryers
was recently reduced from 20% to 10% as a result of recent legislative action.
Emissions of nitrous oxides, sulfur dioxide, volatile organic compounds, and
particulate matter may also present issues for some sites seeking operating
permits. Moreover, thermal drying of combustible particles of coal can present
safety hazards resulting from accidental fires or dust and gas explosions.
The development of an innovative, efficient, and low-cost technology for
removing moisture from fine coal is an important need for the coal prepara-
tion industry. In light of this necessity, a novel thermal-mechanical dewatering
process known as Nano Drying Technology (NDT) has recently been devel-
oped for the coal preparation industry. In the current study, an experimental
test program was undertaken to evaluate the dewatering performance of the
NDT process. This chapter provides a brief description of the new dewatering
technology and presents experimental results obtained from recent bench- and
pilot-scale test programs.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
348 Moisture Reduction and Special Topics
used for extracting moisture from airborne, aerosol, and liquid environments.
Molecular sieves contain pores of a precise and uniform size, typically in the
range of 3 to 10 angstroms (Ramakrishna et al. 2011). These pores are large
enough to draw in and adsorb water molecules, but small enough to prevent
any of the fine coal particles from entering the sieves. Some molecular sieves can
adsorb up to 42% of their weight in water (Bland et al. 2011). Molecular sieves
are used in the drying process because these are reusable after the absorbed
water is removed from the sieves by heating.
Molecular sieves often consist of aluminosilicate minerals, clays, porous
glasses, microporous charcoals, zeolites, active carbon, or synthetic compounds
that have open structures through or into which small molecules such as nitro-
gen and water can diffuse (Breck 1964). When the molecular sieves are mixed
with wet coal fines, these sieves quickly draw water away from the wet solids.
To maximize surface contact between molecular sieves and coal particles,
the mixture is contacted/mixed/agitated for a short period of time (typically
2–6 minutes depending on the characteristics of the coal feed and types of
materials utilized). After contacting, the molecular sieves are recovered from
the dry coal by simple screening given that the sieves are substantially larger in
size than the topsize of the dried coal particles. When the separation occurs, the
remaining coal particles have a substantially reduced moisture content, which
can reach low single-digit values regardless of coal particle size. The molecular
sieves are then regenerated by simple heating to drive off moisture and are
recycled back through the process. It is important to note that the regenera-
tion occurs after the deeply dewatered coal particles have been removed (i.e.,
no portion of the coal is ever subjected to heating). Consequently, this process
is considered by the inventors to be an advanced dewatering process and not a
thermal drying process, which offers many advantages in terms of operational
cost and environmental compliance.
BENCH-SCALE TESTING
Experimental Procedure
A bench-scale experimental test program was performed to evaluate the
performance of the NDT process in removing water from fine coal. For all
experimental tests, the wet feed sample consisted of either 0.6 mm or 0.15 mm
topsize clean metallurgical coal (filter cake) collected from an industrial plant.
During testing, a weighted sample of as-received fine feed coal was mixed with
a predetermined weight of molecular sieves. The mixture was then contacted
together in a small bench-scale rotary mixer for a defined period of time (see
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Nano Drying Technology: A New Approach 349
Dry high-moisture
Separate dewatered Reuse molecular
coal fines by
coal fines from sieves after thermal
combining with
molecular sieves. regeneration.
molecular sieves.
Figure 2). After contacting, the mixture of molecular sieves and coal fines
was separated by using a laboratory sieve. The dewatered coal particles passed
through the sieve and were collected as an underflow product, whereas the
molecular sieves were retained on top of the sieve and were collected as an
overflow product. After separation, the coal particles and molecular sieves were
individually weighed and the reduction in the percentage moisture of the coal
sample was calculated. The last step in the experimental procedure was drying
the molecular sieves. To speed the regeneration process, a microwave oven
was used to evaporate the moisture held in the pores of the molecular sieves.
The regenerated molecular sieves were then reused in the testing program. No
significant difference was observed in the effectiveness of the moisture removal
using either newly manufactured or regenerated molecular sieves.
Five independent “groups” of statistically designed bench-scale experiments
were performed using the NDT process (Table 1). The type (size) of molecu-
lar sieves and weight of coal sample was kept constant for each experimental
group, while the weight of molecular sieves and time of contact were varied
over a range of predetermined values as dictated by the statistical parametric
test matrix. Duplicate test runs (a minimum of three to four) were conducted
at each test point to assess the degree of variability and level of reproducibility
in the test data. The first group of tests (Group A) were comprised of explor-
atory tests designed to identify the suitable ranges of experimental conditions
for testing. This group of test runs involved the processing of five batches of
sample with eight experimental test runs per batch. Groups B and C consisted
of two sets of central composite designs of 39 tests each (15 central point
tests). These groups were identical except for the range of variables examined.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
350 Moisture Reduction and Special Topics
Groups D and E were conducted using a different type (size) of molecular sieve.
The test matrix for Group D consisted of a uniform grid with four batches of
experiments involving 12 test runs each, while Group E consisted of a central
composite design encompassing a single batch of 52 test runs (20 central point
tests). After completing each test matrix, the data were evaluated using standard
statistical techniques.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Nano Drying Technology: A New Approach 351
Batch Product Contact Product Contact Product Contact Product Contact Product Contact
Number Moisture, Time, Moisture, Time, Moisture, Time, Moisture, Time, Moisture, Time,
% min % min % min % min % min
10 — — — — — — 8.91 5 — —
11 — — — — — — 8.34 5 — —
12 — — — — — — 7.81 5 — —
Figure 3 shows the central composite text matrix used in the Group B
test program on the 0.6 mm × 0 feed. As discussed previously, a total of 39
individual test runs were performed in this group using Type I media. The tests
included nine combinations of experiments based on contact time and media
factor. The media factor is a dimensionless number representing the relative
amounts of coal and molecular sieves used in the test run (i.e., a larger media
factor represents a greater addition of feed coal relative to sieve weight, whereas
a smaller number represents less coal relative to sieve weight). The central test
conducted at 4 minutes of contact time and media factor of approximately 0.3
was randomly repeated 15 times throughout the test matrix to evaluate the sta-
tistical reproducibility of the process. Also, each of the satellite tests conducted
around the central test was performed in triplicate to further evaluate the data
reproducibility and to assist in the identification of statistical outliers. All test
runs performed in Group B utilized a constant sample weight of as-received
feed coal. The average moisture content of the feed coal samples used in this
group of tests was 21.9±0.11%.
As shown in Figure 3, all but one of the test runs conducted for the
Group B test matrix gave single-digit moistures in the final 0.6 mm × 0 prod-
uct. The product moistures decreased with either an increase in contact time or
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
352 Moisture Reduction and Special Topics
1.0
3 tests 3 tests
10.23% 8.85%
(0.30) 3 tests (0.55)
9.61%
(0.06)
0.5 3 tests 3 tests
9.84% 8.18%
(0.37) (0.41)
15 tests
8.90%
Media Factor
(0.14)
0.0 3 tests 3 tests
8.45% 7.48%
(1.05) (0.01)
3 tests
7.57%
(0.32)
–0.5
Legend
Number of Tests
Average Product Moisture
(Standard Deviation)
Feed Moisture = 21.9±0.11%
–1.0
1 2 3 4 5 6 7
Contact Time, min
Figure 3 Parametric test matrix and performance data for Group B (0.6 mm × 0 size
feed)
a decrease in media factor (i.e., less coal per unit weight of sieve media). The
standard deviations for each set of conditions varied from a low of 0.01 to a
high of 1.05, which indicated that the data were generally reproducible. In fact,
as shown in Figure 4, the 15 replicate tests conducted at the central point of the
test matrix (i.e., 4 minute contact time and 0.3 media factor) showed little vari-
ability in the product moisture despite significant variations in the feed mois-
ture. The average moisture content for the feed sample used in the 15 central
point tests was 21.8±0.16% with a standard deviation of 0.90. After contacting
with the molecular sieves, the product moisture dropped to an average value
of 8.90±0.02% with a standard deviation of 0.14. The very small confidence
interval and low standard deviation values associated with the data obtained
for the dewatered product indicate that a high degree of reproducibility can be
achieved using the bench-scale version of the NDT process.
A similar trend in moisture removal was observed for the tests conducted
for Group E having a topsize of 0.15 mm. These experiments were conducted
using Type II molecular sieves over a similar range of contact times and a
lower range of media factors. Each satellite test conducted around the cen-
tral test point was repeated four times to assist in identifying outliers and
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Nano Drying Technology: A New Approach 353
25
20 Average = 21.8±0.16%
Std. Dev. = 0.90
15
Moisture, %
10
Average = 8.90±0.02%
Std. Dev. = 0.14
5
0
0 10 20 30 40
Replicate Test Number
Figure 4 Replicate tests conducted at the central test condition for Group B
(0.6 mm × 0 size feed)
evaluating reproducibility. The central test point, which involved a contact time
of 3.5 minutes and media factor of –0.63, was repeated 20 times in random
order throughout the test matrix. For this particular group of tests, the aver-
age moisture contents of the as-received 0.15 mm × 0 feed was 26.2±0.10%.
After contacting with the molecular sieves, the 0.15 mm × 0 product moistures
were reduced to single-digit values for all tests conducted at contact times of
3.5 minutes or longer (see Figure 5). The lowest product moisture content of
6.38% was achieved for the longest contact time of 4.9 minutes. Tests con-
ducted with contact times less than 3.5 minutes did not achieve single-digit
moistures, but at 10.2%–10.9% moisture they were not far from breaking this
meaningful barrier.
One noteworthy difference in the Group E test series was the greater
degree of scatter in the experimental data. Standard deviation values greater
than 1 were observed for the vast majority of the test points and a value as high
as 4.67 was obtained for one of the satellite tests. The increased data scatter
was clearly observed in the 20 replicate tests conducted at the central test point
(see Figure 6). The confidence interval of ±0.19 obtained for the Group E set
of replicate tests was considerably larger than the ±0.02 value obtained for the
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
354 Moisture Reduction and Special Topics
0.0
Legend Feed Moisture = 26.2±0.10%
Number of Tests
Average Product Moisture
(Standard Deviation)
–0.2
4 tests 4 tests
–0.4 10.46% 8.81%
Media Factor
(1.91) (1.32)
4 tests
4 tests 9.55% 4 tests
10.25% (0.53) 6.38%
–0.6 (0.60) (1.06)
20 tests
8.43%
(1.26)
4 tests 4 tests
–0.8 10.95% 6.88%
(4.67) 4 tests (1.56)
8.74%
(1.49)
–1.0
1 2 3 4 5 6 7
Contact Time, min
Figure 5 Parametric test matrix and performance data for Group E (0.15 mm × 0 feed)
Group B tests. Because the moisture of the feed was not determined before
each bench-scale test, the standard deviations for both the feed and product
moistures would be expected to agree, which in fact they do (i.e., 1.29 standard
deviation for the feed moisture and 1.26 standard deviation for the product
moisture). A refinement for future testing would be to determine the moisture
of each feed sample, then calculate the correct media ratio before conducting
the bench-scale test. This is in fact how the pilot-plant tests, described in the
following section, were conducted.
P I L O T - S C A L E D E M O N S T R AT I O N
In light of promising bench-scale data, a decision was made to construct a pilot-
scale NDT plant to demonstrate the capabilities of this new technology in
continous mode. The flowsheet for the facility is shown in Figure 7. The com-
pleted facility, which was largely assembled using off-the-shelf components, was
designed with an effective throughput capacity of 1,000 lb/h (0.5 tph). The
self-contained facility included unit operations for handling, contacting, and
separating the coal and media. An advanced gas-fired dryer was used to regen-
erate the molecular sieves such that the entire process operated in a closed-
circuit loop. The prototype facility was designed, constructed, and successfully
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Nano Drying Technology: A New Approach 355
30
25
Average = 26.2±0.19%
Std. Dev. = 1.29
20
Moisture, %
15
10
5 Average = 8.43±0.19%
Std. Dev. = 1.26
0
0 10 20 30 40
Replicate Test Number
Figure 6 Replicate tests conducted at the central test condition for Group E
(0.15 mm × 0 size feed)
Feed Coal/Sieve
Coal Contactor
Water Sieve
Vapor Thermal Dryer
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
356 Moisture Reduction and Special Topics
DISCUSSION
The removal of unwanted moisture from fine coal has historically been con-
sidered one of the most challenging technical problems in the coal preparation
industry. The NDT process was developed specifically to address this issue
by providing effective moisture removals, efficient energy utilization, and
enhanced environmental performance. Experimental data collected from both
bench- and pilot-scale operations show that single-digit moisture values can be
readily achieved from fine coal feeds containing 30% moisture or more. The
process is highly flexible in that the product moisture can be “dialed in” by
varying contacting time and coal-to-seive media loadings. Also, unlike existing
mechanical processes, the product moisture from the NDT process is largely
independent of the particle size distribution of the feed stream.
The removal of moisture in the NDT process occurs at ambient tempera-
ture. As such, the coal particles are never subjected to high temperatures, which
greatly reduces the emissions of criteria pollutants that are normally associated
with conventional coal drying systems. Recent estimates by an environmental
consulting group indicate that emission reductions as large as 90% or more
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Nano Drying Technology: A New Approach 357
1,200
PM
CO
NOx
1,000 SO2
VM
800
Criteria Pollutants, t/a
600
400
200
0
Thermal Dryer NDT
Figure 8 Comparison of air emissions for thermal drying and the NDT process
when compared to a thermal dryer are possible using the NDT process. The
emission projections from one such case study is shown in Figure 8. In this case,
emissions of volatile matter (VM), sulfur dioxide (SO2), and particulate matter
(PM) would be essentially eliminated (>99% reduction) using the NDT pro-
cess. Projected emissions of carbon dioxide (CO2) and nitrous oxides (NOx)
would be reduced by 91% and 84%, respectively. It is particularly important to
note that the projected total emissions of 59.4 t/a (tons per annum) of criteria
pollutants is likely to be less than the threshold value that would trigger the
need for a Title V Air Quality permit in many states. For example, no such per-
mit would be required in West Virginia because the threshold value is 100 t/a
of criteria pollutants. The NDT process also requires no added chemicals and
generates no other by-products that could potentially be released into the
environment.
Finally, the NDT process is very efficient in terms of energy utilization.
Given that only the molecular sieves are thermally dried, the drying step can
be fully optimized in the absence of coal-imposed contraints associated with
dryer temperature levels, gas-solid contacting systems, and coal dust explosions.
As such, the system provides the highest possible energy efficiency at the lowest
possible fuel cost. Because the process treats only the fine coal fraction, which
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
358 Moisture Reduction and Special Topics
120
Fuel
Chemicals
Drying Media
100 Electricity
Maintenance
Labor
80
Relative Cost, %
60
40
20
0
Thermal Dryer NDT
Figure 9 Comparison of relative costs for thermal drying and the NDT process
is generally between 10% and 15% of the total clean coal product (and not the
entire clean coal product treated by conventional thermal dryers), the required
footprint for the facility is only a fraction of that demanded by a large-scale
coal thermal dryer. Also, because of fewer operational complexities, significant
cost savings are also expected for ancillary items such as electricity, chemicals,
maintenance, and labor. Cost estimates conducted in cooperation with a com-
mercial engineering firm are plotted in Figure 9. Although such economic
calculations tend to be site specific, the costing figures for this site do suggest
a relative operating cost of less than half of that required to operate a conven-
tional thermal dryer.
SUMMARY
The removal of surface moisture from fine coal has been a long-standing prob-
lem in the coal industry. To address this need, an innovative process based on
nanotechnology has been developed. Bench-scale studies indicate that NDT is
an effective method for coal drying. The NDT process can effectively dewater
fine (1 mm × 0) coal from slightly more than 30% surface moisture to single-
digit values. Test data obtained using a pilot-scale NDT plant further validated
this impressive capability using a continous prototoype facility. It was also
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Nano Drying Technology: A New Approach 359
observed that, unlike existing fine coal dewatering processes, the performance
of the NDT process is not dictated or constrained by particle size (i.e., it works
equally well on 1 mm × 0 coal as it does on 325 mesh × 0 coal). The NDT pro-
cess overcomes problems associated with other techniques for fine coal drying
because dewatering occurs at ambient temperature and low airflow. Only the
molecular sieves have to be dried, which reduces energy. Moreover, this process
requires no chemicals, produces no damaging contaminants, and has a very
small installed footprint and environmental impact.
REFERENCES
Bland, R.W., Harsh, P., Hurley, M., Jones, A.K., and Sikka, V.K. 2011. U.S. Patent
Application Publication, Pub. No. US 2011/0078917 A1 Pub. Date Apr. 7.
Breck, D.W. 1964. Crystalline molecular sieves. J. Educ. 41(12):678–689.
Keles, S. 2010. Fine coal dewatering using hyperbaric filter centrifugation. Ph.D. dis-
sertation, Virginia Polytechnic Institute and State University. Blacksburg, VA. pp.
4–18.
Le Roux, M., Campbell, Q.P., Watermeyer, M.S., and de Oliveira, S. 2005. The optimi-
zation of an improved method of fine coal dewatering. Miner. Eng. 18(9):931–934.
Orr, F.M. 2002. Coal Waste Impoundments: Risks, Responses and Alternatives. Washing-
ton, DC: National Research Council.
Osborne, D.G. 1988. Solid-liquid separation. In Coal Preparation Technology. Vol. 1.
London: Graham and Trotman. pp. 478–542.
Ramakrishna, S., Ma, Z., and Matsuura, T. 2011. Chapter 1. In Polymer Membranes in
Biotechnology. London: Imperial College Press.
Wills, B.A., and Napier-Munn, T.J. 2006. Chapter 15. In Wills Mineral Processing Tech-
nology. 7th ed. Amsterdam: Butterworth-Heinemann.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of a Fine Coal
Recovery Operation at the
Centralia Mine Coal Slurry
Impoundment Structures
C. David Henry
ABSTRACT
Coalview Recovery Group, LLC, is in the final stages of establishing a fine coal
recovery operation at the TransAlta-Centralia mine located near Centralia, Wash-
ington. The operation will extract coal slurry refuse from three impoundment struc-
tures identified as “Series 3 Ponds” containing an estimated 18 to 20 million tons
of slurry refuse material. The coal slurry refuse will be delivered to the Coalview
Recovery Group, LLC, fine coal recovery plant by use of a 200-tph all-electric
hydraulic dredge. The fine coal plant will process the nominal +74 µm coal slurry
refuse and produce a fine coal product to be transported to the adjacent TransAlta-
Centralia coal-fired power plant. The slurry refuse developed by Coalview Recovery
Group, LLC, fine coal recovery plant will be treated through a deep-cone paste
thickener and pumped to a below-grade, incised impoundment for final deposition.
PROJECT DESCRIPTION
Coalview Recovery Group, LLC (CRG), will enter into a contract with
TransAlta-Centralia Mining to develop a fine coal recovery operation at their
mine site located near Centralia, Washington. The purpose of this project is to
• Evacuate the majority of coal slurry refuse (CSR) from the three
impoundment structures;
• Process the CSR through the CRG fine coal recovery (FCR) plant to
produce a fine coal product;
361
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
362 Moisture Reduction and Special Topics
• Transport the fine coal product from the FCR plant to the TransAlta-
Centralia coal-fired power plant within a mile of the recovery opera-
tion; and
• Treat the slurry refuse from the CRG-FCR plant through a deep-cone
paste thickener and pump the resulting paste into an on-site, below-
grade, incised impoundment for final deposition.
Impoundment Structures
The CSR, contained within the impoundment structures, was deposited from
the TransAlta-Centralia mine coal preparation plant and operated from 1971
to 2006. The impoundment structures to be evacuated are part of a system of
five structures known as the “Series 3 Ponds.” The impoundment structures
containing CSR are identified as Pond 3B, Pond 3C, and Pond 3D and contain
3.5 million, 3.9 million, and 10.5 million tons of solids, respectively. Pond 3E
or the North Hanaford Pit, is the impoundment structure that will contain
the slurry refuse from the FCR plant containing an estimated volume of 59.1
× 106 yd3 (45.2 × 106 m3). Pond 3A is included in the Series 3 Ponds but will
not be included in the FCR project. Pond 3A is located at the toe of Pond 3B
and receives overflow water from Pond 3B before discharge into the regulated
NPDES (National Pollutant Discharge Elimination System) outflow point.
Pond 3A has an approximate surface area of 70 acres (28.3 ha) with a total
depth of about 50 ft (15.2 m). Pond 3B was the initial impoundment developed
by the Centralia mine for the CSR disposal and filled to capacity in 1986. The
3B pond covers about 112 acres (45.3 ha) at a maximum depth of about 130 ft
(39.6 m) and contains approximately 8.3 × 106 yd3 (6.3 × 106 m3) of CSR and
3.3 × 106 yd3 (2.5 × 106 m3) of water. Pond 3C covers about 91 acres (36.8 ha)
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of a Fine Coal Recovery Operation 363
Top of
Embankment First Downstream
Embankment Addition
Coal Slurry Refuse
Second Downstream
Embankment Addition
Starter
Embankment
Downstream Valle
y Flo
or C
ontour
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
364 Moisture Reduction and Special Topics
• Sweet zone: This material is most often considered the “best” for
recovery. It contains the most lower-ash coal and is located at the
outermost areas of the delta zone and within the slurry water interface.
• Far discharge: This is the finest sized with the lowest density; the mate-
rial in this zone consists of clays/slimes and may have considerable
amounts of ultrafine coal.
Sampling and analysis results indicating the discharge point(s) for the vari-
ous impoundment structures are listed as follows:
• Pond 3B is discharged from the upper valley area at the southeastern
and southern end of the impoundment.
• Pond 3C is discharged from potentially two to three points along the
southwestern side of the impoundment.
• Pond 3D is discharged from one point at the northwest corner of the
impoundment.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of a Fine Coal Recovery Operation 365
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
366 Moisture Reduction and Special Topics
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of a Fine Coal Recovery Operation 367
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
368 Moisture Reduction and Special Topics
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of a Fine Coal Recovery Operation 369
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
370 Moisture Reduction and Special Topics
Notes
1. The water level of the dredge hole should be dropped to a
depth of at least 20′ to 40′ below the surface of the slurry.
2. Slurry material should be dredged at a depth of at least 20′.
3. A water cannon mounted on the dredge and/or on the bank
should be used to saturate and wash the exposed slurry material.
4. The saturation and washing actions will cause the following effects:
a. the failure and wash of the exposed slurry face and the
creation of a beach area providing a safety zone between
the dredge and exposed slurry face; this beach area should
be maintained at no less then 20′ and
b. additional weight to the exposed slurry material above the
water level, enhancing the downward action and movement
of the slurry material toward the cutterhead.
5. The slewing arc should be maintained at the longest allowable
Top of Slurry length and as specific site conditions allow.
Beach Area
20′ to 100′+
20’ to 40’
Water Cannon
Top of Water
Line of Movement/Force
of Slurry Material to
Dredge Cutterhead
20’ ±
Figure 4 CRG general dredging plan for the removal of coal slurry material
P R O J E C T O P E R AT I O N S
The recovery operation is anticipated to begin processing during the third quar-
ter of 2013. The FCR facility will operate 24 hours per day, 7 days per week for
50 weeks per year and deliver a minimum of 200 tph of CSR to the processing/
refuse treatment system. The facility is expected to operate for 12 to 15 years
with earliest completion expected in 2025. The average anticipated clean coal
yield is estimated to be 24% with an estimated 1,428,000 tons of CSR for pro-
cessing and producing 342,000 tons of fine coal product to be delivered to the
adjacent TCPP.
Mining/Dredging Plan
The recovery operation will begin in Pond 3C, then move to Pond 3B and Pond
3D, respectively, as the operations advance. All slurry refuse developed from
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of a Fine Coal Recovery Operation 371
the CRG-FCR plant will be deposited into Pond 3E. The dredging units and
pipelines will be relocated from impoundment to impoundment as the FCR
operation proceeds. The general dredging plan is depicted in Figure 4.
The mining/dredging plan will be developed by dividing each impound-
ment into a series of dredge cells. The dredge cells are established around
the Vibracore samples extracted for analysis, with the sample point being in
the center of the cell area. The quality of each dredge cell will be determined
by interpolating the quality between each sample point and establishing the
overall average quality for the dredge cell. As the dredging continues to greater
depths of the impoundment, added Vibracore samples will be extracted and
new dredge cells/quality established. This pattern will continue to the maxi-
mum depths of the impoundment structures. The layout of the dredge cells in
Pond 3B is shown in Figure 5.
The operation will be a two-dredge system; the primary dredge will operate
in dredging cells designated to have adequate recovery rates, feeding the FCR
facility a minimum of 200 tph, and the secondary dredge will deliver no/low-
yield CSR directly to the static/paste thickener system at a rate of 40 to 50 tph.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
372 Moisture Reduction and Special Topics
long periods of time. The Centralia mine treated the CSR from their process
in a static thickener, attempting to cause the clay material to settle out, with
little effect. The clay that was discharged from the thickener underflow into
the impoundment structures remained suspended as long as slurry material was
pumped into the impoundment structure. The clarified water from the static
thickener also contained substantial amounts of the clay, creating problems
within the Centralia preparation process. As the mining and preparation activi-
ties continued, the concentration of the clay in the impoundment and clarified
water continued to increase, causing problems in the processing system.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of a Fine Coal Recovery Operation 373
As the CSR material was discharged into the impoundment structures, the
majority of more dense materials and larger particle coal settled and extended
out and under the suspended clay materials. When the impoundment struc-
tures reached capacity and/or the mining and preparation operations ceased,
the clay slowly settled and covered the underlying materials, developing an
“overburden” on the more coarse and higher recoverable reserves. This is
confirmed in the evaluation of a series of deep core samples taken from the
impoundment structures and shown in one of the deep core sample summaries
shown in Table 4. The dredge cells of no/low yield will be identified and the
secondary dredge will pump this material directly to the static/paste thickener
to be developed into a paste for final deposition into Pond 3E.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
374 Moisture Reduction and Special Topics
PROJECT SUMMARY
The development of recovery operations at slurry waste impoundment struc-
tures provides a variety of benefits. These benefits include but are not limited to
• Regaining minerals or energy from the waste facility and returning to
end users as a usable product,
• Reduction of the quantity of slurry materials and water volumes
within the impoundment structure or evacuating all materials and
relocating them to another area,
• Decreasing the reclamation costs and enhancing the final reclamation
plan, and
• Eliminating the Mine Safety and Health Administration hazard risk
ranking of the impoundment structures.
The installation of the fine coal recovery system at the TransAlta-Centralia
mine site will incorporate all the benefits previously listed. CRG will process an
estimated 16.2 million tons of raw coal slurry to produce 3.42 million tons of
fine coal product to be transported directly to the adjacent TransAlta-Centralia
complex, a 1,404-MW two-unit coal-fired power plant. The coal slurry refuse
resulting from the CRG-FCR plant will process and treat this refuse through
the two-stage static-paste thickener system and develop a more stable paste
material for final deposition into Pond 3E, a final-cut, below-grade, incised
impoundment. An additional 1.86 million tons of no/low-yield material will
bypass the CRG-FCR plant into the thickener system for treatment before final
deposition into Pond 3E.
This project will evacuate nearly all the coal slurry refuse and water from
the three downstream-constructed impoundment structures—Ponds 3B, 3C,
and 3D—and deposit the more stable paste into Pond 3E. This recovery opera-
tion will reduce the reclamation cost to TransAlta-Centralia mine by
• Eliminating the requirement to create a positive drainage from the
upstream portion of the impoundment, across the slurry surface to
discharge through the impoundment berm;
• Reducing the requirement of soil cover on the slurry surface to create
a stable surface area and one that will sustain a vegetative growth of
natural and native flora; and
• Doing away with the maintenance and repair of soil cover erosion and
exposure of the slurry subsurface.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Development of a Fine Coal Recovery Operation 375
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Pressurized Fluidized-Bed
Combustion Technology for
Fine Coal Utilization*
Deborah A. Kosmack, Michael Fenger,
and Esko Polvi
ABSTRACT
Waste coal, otherwise known as “culm” or “gob,” has been stored in large quantities
throughout the coalfields of Pennsylvania and other Appalachian states. Some of
this legacy waste material and much of the continuing production of waste coal is
in the form of wet fine particles stored in impoundments. Pressurized fluidized-
bed combustion (PFBC) technology can use these impounded fines for highly
efficient electric power generation. PFBC plants in the 100–400 MWe (megawatt,
electric) size range have been in successful commercial operation since 1990 with
demonstrated efficiencies of more than 40% higher heating value and availabili-
ties of 85%–90%. A small footprint and inherently low emissions make PFBC
combined-cycle technology an attractive choice for repowering dated and difficult-
to-permit pulverized coal power plants. The PFBC process is very robust and it can
efficiently burn a wide range of fuels. It is well suited for high-moisture fine coals
and it has been tested with a variety of fuel mixes with low and high Btu values,
including a wet waste coal/biomass mix, without significant detrimental impact
on the combustion behavior or emissions. To generate the project-specific data
necessary for designing commercial power plants, PFBC Environmental Energy
Technology, Inc., and Consol Energy, Inc., operate a 1-MWt (megawatt, thermal)
PFBC pilot plant. This chapter will describe the process test facility and selected
results of its operation on fine wet waste coal material.
* All rights, title, and interest to this work, including the copyrights, belong to
PFBC-Environmental Energy Technology, Inc., and Consol Energy, Inc. PFBC-
Environmental Energy Technology, Inc., and Consol Energy, Inc., have given
permission to SME, Inc., to include this work in this book.
377
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
378 Moisture Reduction and Special Topics
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Pressurized Fluidized-Bed Combustion Technology 379
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
380 Moisture Reduction and Special Topics
P F B C F O R F I N E C O A L U T I L I Z AT I O N
The fine waste coal from active coal preparation plants represents a significant
resource. Given that the cost of cleaning fine coal in the preparation plant can
be three to four times higher than the cost of cleaning larger size coal, the fine
coal and refuse are often discarded to a waste coal impoundment as slurry.
The discarded wet fines often contain 50% or more coal. These fines can be
dewatered to 25%–30% moisture to produce a pumpable paste, which can be
injected in a fluidized-bed combustor. Fluidized-bed combustors are inherently
fuel flexible. High pressure (12 to 16 bars) increases the density of the air/
oxygen relative to fuel particles. The turbulence, the deepness of the bed, and
the low gas velocity (leading to longer residence time) all contribute to nearly
complete oxidation of carbon in the fuel. Wet and high-moisture fuels create
additional water vapor mass flow through the gas turbine expander, increasing
the power output. Therefore, cycle efficiency is only slightly reduced by wet
fuels, unlike pulverized-coal-fired plants or atmospheric fluidized-bed units
where the water vapor is simply exhausted. To reduce the amount of waste coal
pumped to impoundments from active coal preparation plants, the PFBC plant
can be integrated with a coal preparation plant.
Pennsylvania Department of Environmental Protection records show that
the coal refuse piles in Pennsylvania which were abandoned prior to 1978
included 819 separate sites in 35 counties. Those sites contain an estimated
170 million tons of material that covers 8,021 acres of surface land. These
figures include only those refuse piles abandoned before 1978 yet exclude
many others, such as a 40-to-70-million-ton pile in Washington County,
which was still active in 1978. Some of these refuse sites contain substantial
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Pressurized Fluidized-Bed Combustion Technology 381
concentrations of coal, and the material may have a heating value of 8,000 Btu/lb
or more; others have much lower energy values, about 2,000 Btu/lb. When
compared to coal on an energy-content basis (e.g., lb/MBtu), coal refuse often
contains higher concentrations of sulfur, mercury, and other deleterious ele-
ments. Thus, whatever technology is employed to consume the coal refuse must
be highly effective in controlling emissions of pollutants and must be able to
burn low-grade fuel of highly variable quality.
P F B C P R O C E S S T E S T FA C I L I T Y
PFBC-EET and Consol Energy joined together in 2006 to
• Design and reconstruct a 1-MWt input PFBC Process Test Facility
(PTF) in South Park, Pennsylvania (pressure vessel and fluidized-bed
combustor relocated from Alstom’s PTF in Finspong, Sweden);
• Demonstrate the operation of PFBC technology using fine wet waste
coal as the fuel; and
• Evaluate performance of PFBC on waste coal/biomass with CO2
capture.
The purpose of the PTF pilot-plant research test program is to provide the
necessary fluidized-bed combustion and emissions data needed to design and
build commercial-scale pressurized fluidized-bed combined gas-steam cycle
generating units that will operate on fine waste coal. The operating pilot-scale
PTF at Consol Energy Research & Development was constructed to conduct
tests to determine
• Fuel properties and sorbent selection;
• Fluidized-bed temperatures and heat transfer;
• Flue gas emissions and emissions sensitivity;
• Bed ash and fly ash distribution, handling, and utilization samples; and
• Ash samples for lab analysis and utilization tests.
The data from the PTF testing will be used to prepare
• Operations and maintenance projections;
• A fuel impact model;
• A fluid dynamic model for boiler, combustor, and cyclone design;
• Operator training;
• Risk assessment and mitigation for equipment design; and
• Define scope for full-scale plant design and cost estimates.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
382 Moisture Reduction and Special Topics
The pilot plant consists of a wet and dry fuel handling system, pressurized
combustion vessel, and ash handling system. Coal and limestone sorbent are
mixed with water and fed as a paste into a pressurized fluidized-bed combus-
tor. The combustion process occurs in a fluidized bed that is characterized by
a low gas velocity and high residence time, which provides high combustion
efficiency and low pollution emissions.
C L E A N C O A L A N D F I N E W A S T E C O A L P T F T E S T D ATA
C O M PA R I S O N
A series of test runs were conducted to commission and calibrate; the PTF on
three different coals. The coals included a Bailey mine clean coal; a coal (Pol-
ish) used at the Vartan PFBC plant in Stockholm, Sweden; and a fine waste
coal from a Consol preparation plant. The test runs were designated as follows:
• Run 14, Bailey clean coal with Maple Grove Quarry (MGQ) dolomite
limestone sorbent
• Run 15, Vartan coal with Vartan dolomite limestone sorbent
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Pressurized Fluidized-Bed Combustion Technology 383
Figure 3 PTF pressure vessel, bed vessel, and cyclones process diagram
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
384 Moisture Reduction and Special Topics
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Pressurized Fluidized-Bed Combustion Technology 385
900
880
860
Temperature, °C
840
820
800
760
#14 Bailey CC
#14 Bailey CC
#14 Bailey CC
#14 Bailey CC
#15 Vartan
#16 Vartan
#16 Vartan
#16 Vartan
#20b W Coal
#20b W Coal
#20b W Coal
#20c W Coal
#20c W Coal
#20c W Coal
#20c W Coal
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
386 Moisture Reduction and Special Topics
0.40 8
SO2 lb/MBtu
0.35 NOx lb/MBtu 7
O2 %
0.30 6
NOx or SO2, lb/MBtu
0.25 5
O2, %
0.20 4
0.15 3
0.10 2
0.05 1
0.00 0
#14 Bailey CC
#14 Bailey CC
#14 Bailey CC
#14 Bailey CC
#15 Vartan
#16 Vartan
#16 Vartan
#16 Vartan
#20b W Coal
#20b W Coal
#20b W Coal
#20c W Coal
#20c W Coal
#20c W Coal
#20c W Coal
sintering or cyclone ash deposits during any of the test runs. This indicated
that the temperatures were not excessive. Further commercial test programs
will verify this temperature pattern. The temperature pattern will influence the
design of the boiler, cyclones, and gas turbine expander.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Pressurized Fluidized-Bed Combustion Technology 387
100
C2 Ash %
90 C1 Ash %
26 25 Bed Ash %
31
80 34
48 45 47 45
51 50
70 57
60
78
81 79 72
Sulfur, %
50
40
72 73
30 64 68
51 53 52 54
47 49
20 41
10 22 19
16 17
0
#14 Bailey CC
#14 Bailey CC
#14 Bailey CC
#14 Bailey CC
#15 Vartan
#16 Vartan
#16 Vartan
#16 Vartan
#20b W Coal
#20b W Coal
#20b W Coal
#20c W Coal
#20c W Coal
#20c W Coal
#20c W Coal
of sulfur flow in the three ashes collected during separate test periods of each
test run. Test run 14 shows that most of the sulfur was removed with the bed
ash. Test runs 15 and 16 results show the effect of a very high calcium/sulfur
(Ca/S) ratio. Test run 20 at 26% sorbent in the fuel shows lower sulfur in the
bed ash compared to test run 14 at 13% sorbent in the fuel. The lower sulfur
flow during test run 20 was most likely due to the larger amount of sorbent in
the paste fuel (coal/sorbent). The increased flow of sorbent into the fluidized
bed will increase the production of bed ash, which reduces the residence time
of the sorbent in the bed and thereby reduces sulfur absorption. The sorbent
properties, particle sizing, and particle attrition in the bed will influence the
distribution of the sulfur in the three ashes.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
388 Moisture Reduction and Special Topics
12
7
Ca/S ratio
0
#14 Bailey CC
#14 Bailey CC
#14 Bailey CC
#14 Bailey CC
#15 Vartan
#16 Vartan
#16 Vartan
#16 Vartan
#20b W Coal
#20b W Coal
#20b W Coal
#20c W Coal
#20c W Coal
#20c W Coal
#20c W Coal
Figure 8 Calcium/Sulfur ratio in C1, C2, and bed ash (1.0 is optimum)
the excess amount of calcium needed to sustain the bed level and the very low
sulfur in the Vartan coal. Test run 20 shows poorer calcium utilization com-
pared to run 14, most likely due to the larger amount of sorbent in the paste
fuel, which reduced the residence time of the sorbent in the bed and raised the
bed ash removal rate. Adjustment of the limestone sorbent particle sizes may
be necessary to improve calcium utilization and thereby reduce the amount of
sorbent required. The Ca/S ratios of 1.7 and 1.8 in the fuel during test runs 14
and 20 could potentially be lowered and still achieve acceptable sulfur removal.
The third test period during run 14 shows an advantageous condition where all
three ashes are at the lowest ratios.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Pressurized Fluidized-Bed Combustion Technology 389
60
58
56
54
52
50
%
48
46
44
42
40
#14 Bailey CC
#14 Bailey CC
#14 Bailey CC
#14 Bailey CC
#15 Vartan
#16 Vartan
#16 Vartan
#16 Vartan
#20b W Coal
#20b W Coal
#20b W Coal
#20c W Coal
#20c W Coal
#20c W Coal
#20c W Coal
Figure 9 Tube bundle heat removal—% of input heat from coal
CONCLUSIONS
Fine coal utilization involves many processing steps, starting at the coal prepa-
ration plant or impoundment and finishing with combustion at a power plant
boiler. Based on the early results of test runs at the PTF, fine wet waste coal can
be burned efficiently and cleanly using PFBC technology. The major conclu-
sions from the current work are as follows:
• The fine waste coal was successfully burned in the 1-MWt PFBC PTF
in South Park, Pennsylvania. The low fluidized-bed superficial gas
velocity of 1 m/s is adequate to maintain combustion efficiency of
98% or more.
• The sorbents selected for the test runs performed well, but further
testing of potential sorbent sources and size distributions need to
be completed to improve calcium utilization and thereby reduce the
amount of sorbent required.
• The fine coal and larger amounts of ash do not appear to influence
emissions as compared to the coarser and cleaner coals.
• The amount of fine ash influences the temperature profile in the com-
bustor, which influences the fluidized boiler design.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Index
Agglomeration, 18–19 Belgium, and dense-medium cyclones,
Alabama 140–141
Centribaric centrifuges in capture of fine Billiton Mitsubishi Alliance, 261–277
coal lost to refuse stream (No. 7 Blackwater, Australia, 261
mine), 303–308 high-speed disc filters in coal ultrafines
Jim Walter Resources No. 7 mine processing for coking coal, 261–277
(Brookwood), 303 Bokela Boozer disc filters, 261–277
Anglo American Thermal Coal South Briquetting, 18–19
Africa, 95–98, 120–121 Brookwood, Alabama
briquetting, 100–102, 108–109, Centribaric centrifuges in capture of fine
118–119, 120 coal lost to refuse stream (No. 7
dewatering and filtration, 99–100, mine), 303–308
104–106, 114–117, 120 Jim Walter Resources No. 7 mine, 303
dual cell flotation, 110–114, 119
flotation, 98–99, 102–103, 106–107, Cavitation-tube sparging technology, 195
109–110 Centralia, Washington, 361
multicell flotation, 103 Coalview Recovery Group in processing
Arch Coal, Inc., 67–68 of slurry refuse from TransAlta-
fine pyrite rejection by cyclone and Centralia mine, 361–375
sieve classifiers, compound spiral Centrifugal flow-film concentrators,
concentrators, and reflux classifiers 231–234
(Leer Mining plant, W. Va.), 68–78 Centrifugal fluidized-bed concentrators,
Australia 234–236
clay issues, 126–133 Centrifugal jigs, 236–238
coal grain analysis, 133–134 Centrifuges, 280–281
coal markets, 124–126 Centribaric centrifuges in capture of
and dense-medium cyclones, 145–147 fine coal lost to refuse stream ( Jim
high-speed disc filters in coal ultrafines Walter Resources No. 7 mine,
processing for coking coal, Brookwood, Ala.), 303–308
261–277 Chemical separation, 15–19
mechanical flotation cell usage, 221–224 China
regional coal types, 123–124 and dense-medium cyclones, 147–149
University of Newcastle and Ludowici mechanical flotation cell usage, 224
Australia in development of reflux Coal
classifiers, 159–183 clays in, 59–62
391
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
392 CHALLENGES IN FINE COAL PROCESSING
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Index 393
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
394 CHALLENGES IN FINE COAL PROCESSING
© 2012 by the Society for Mining, Metallurgy, and Exploration. All rights reserved.
Challenges in Challenges in