Upload
trinhnga
View
246
Download
0
Embed Size (px)
Citation preview
The Southern African Institute of Mining and Metallurgy
Platinum 2012
361
T. Nyakudarika, S. Makumbe, C.Rule
FULLY AUTOGENOUS GRINDING AT UNKI MINE CONCENTRATOR – A
CASE OF SUCCESSFUL VALUE ENGINEERING
T. Nyakudarika DRA Mineral Projects (Pty) Ltd
S. Makumbe Anglo American Platinum Limited
C. Rule Anglo American Platinum Limited
Abstract
The Unki Platinum Mines concentrator, employing an MF2 circuit configuration which includes
fully-autogenous primary milling of run-of-mine (ROM) ore, was commissioned in November
2010. The project was Anglo American Platinum’s first PGM concentrator in Zimbabwe and
would form a significant contribution to the group’s platinum production. The project financials
demonstrated that the project was operating-cost sensitive.
After reviewing the metallurgical testwork, mineralogical examinations, and the history of
similar operations, it became apparent that the Unki ore characteristics rendered it amenable
to autogenous milling. The primary ball mill, structurally designed for 30 per cent steel load,
could be operated at the required initial phase throughput with fully-autogenous primary
milling, realizing savings in both power requirements and grinding media consumption. The
original mill sizing had been done with fully-autogenous milling in mind, but allowing for
flexibility and mitigation in case of varied mining fragmentation.
Since attaining steady state, target grinds have been achieved. This paper looks at the Unki ore
characterization programme, design considerations, and the operational and control philosophy
for a successful comminution circuit, and discusses the design to produce a flexible, low opex,
mitigated comminution and overall MF2 flowsheet that is easily expanded in line with the
potential for optimum exploitation of the orebody.
Introduction
The Unki Mines platinum project is the first Anglo American Platinum operation in Zimbabwe to
exploit the rich platinum resources of the Great Dyke. Due to the size of the ore resource,
future expansion of the scale of the Unki operation is an important consideration. The circuit
design considered this aspect carefully. The exploitation strategy adopted for the Unki ore
resource was to start the operation at 120 kt/month and ramp up to 240 kt/month after five
years.
The Southern African Institute of Mining and Metallurgy
Platinum 2012
362
It was imperative for the design to produce a flexible, low operating cost, but fully risk-
mitigated comminution and overall MF2 flowsheet that is easily expanded in line with the
potential for optimum exploitation of the orebody. Figure 1 is the locality map for the Unki
Mine.
Figure 1- Map of Zimbabwe showing the location of Unki on the Great Dyke
A MF2 circuit configuration employing run–of-mine (ROM) primary ball milling, followed by
secondary ball mill regrinds, was initially proposed after a full risk evaluation, based on the
comprehensive testwork programme conducted on borehole cores and bulk samples from the
property. Target grind for primary milling was 35 to 40 per cent passing 75 µm (corresponding
to a P80 of 212 µm), with a final grind of a minimum 80 per cent passing 75 µm. The target
platinum recovery for this circuit was 82 per cent. The financial evaluation of the project
demonstrated that the project was primarily operating-cost sensitive.
The ore characteristics from drop weight tests conducted on geological and metallurgical
samples and the estimated ’as is’ ROM ore size distribution presented an opportunity for
changing the primary milling circuit from ROM primary ball milling to fully autogenous (FAG)
primary milling. An important interpretation for FAG milling utilization is the expected
fragmentation from blasted ore mined by mechanized bord and pillar’ mining. This would result
in a saving in grinding media and power in the primary milling stage.
The Southern African Institute of Mining and Metallurgy
Platinum 2012
363
By suitably sizing the mills, future circuit flexibility could be incorporated in the design for
minimal initial cost.
There was concern regarding this approach, as testwork had previously indicated that the Unki
ore exhibited some degree of bimodality from a perspective of broken ore particle size, and this
could result in critical size build-up in the mill, and as such represented a risk to the success of
autogeneous milling.
The Unki concentrator has achieved the design target of throughput and grind with the
projected savings in grinding media and primary mill power by employing a FAG primary milling
circuit.
The paper discusses the testwork programme, design, and evaluation of the plant performance.
Test work
Extensive metallurgical test work over a prolonged period was conducted to characterize the
orebody and to formulate the plant flow sheet. Laboratory and pilot-plant testwork were
conducted at Geomet (Harare), AARL (Crown Mines), ARC (Anglo Research Centre, Germiston),
Mintek (Randburg), and DML (Anglo American Platinum’s Divisional Metallurgical Laboratory,
Rustenburg). Eurus Mineral Consultants (EMC) validated the laboratory and pilot-plant scale
testwork results using their ’Kincalc’ and ’Supasim’ simulation software. Significant evaluation
of the Unki ore resource commenced in 1992 with testwork conducted on 22 drill cores over
the mining area and six bulk ore samples from a trial mining shaft. Further test work was
continued from the early 2000s up to the time of project execution. The test work from 2002 to
2008 was on more drill cores and two bulk ore samples (2005 and 2008).
Mineralogy
The mineralogical work indicated that the ore was typical of the Great Dyke ores.
Methods of investigation used involved:
• X-ray diffraction (XRD) for comparative semi-quantitative gangue mineral investigation
• Scanning electron microscopy (SEM) and optical microscopy to determine the mass
distribution of sulphides
• Mineral Liberation Analyser (MLA) and QEMSEM for analyses of mineral associations
• X-ray fluorescence (XRF) scans to identify other valuable constituents of the ore samples
• ICP scans to check for variations in composition between the drill core and bulk samples.
The ore has high talc (8 to 12 per cent) and pyroxene (approximately 70 per cent) contents. The
predominant platinum group minerals (PGMs) are sperrylite and Pd-Bi Tellurides. The bulk of
the platinum group elements PGEs occur as free PGMs. The platinum-bearing PGMs are mainly
associated with pyrrhottite (approximately 66 per cent) and lesser amounts (approximately25
per cent) with pentlandite, while the palladium-rich minerals are associated with gangue
(approximately 79 per cent). The gold particles are mostly associated with chalcopyrite1, 2
The Southern African Institute of Mining and Metallurgy
Platinum 2012
364
The base metals sulphides are predominantly pyrrhotite, pentlandite, chalcopyrite, and pyrite.
The average size of the PGMs varies from 5 to 30 µm with a maximum of 90 µm. The base
metal sulphides vary from 5 to 20 µm. The detailed head analysis for the ore at a mining cut of
1.8 m is as follows:
Pt 1.83 g/t, Pd 1.56 g/t, Rh 0.16 g/t, Au 0.30 g/t, Ni 0.21%, and Cu 0.15%.
The mineralogical data formed the basis for optimizing key design parameters and also for
benchmarking against the other Great Dyke operations.
Comminution
AARL, ARC, Mintek, and later DML conducted a series of drop weight test (DWT) and Bond
Work Index (BWI) test work on core and bulk samples of the Unki ore. JKMRC analysed the
Mintek work. Comparison between Unki and other Great Dyke ores was also done. The Unki
ore characteristics are typical of the Great Dyke ore. The ore displays moderately hard range of
resistance to impact breakage and medium abrasion range. Test methods used included:
• Calibrated AARL Rod Mill by AARL
• JKRMC standard procedure for the DWT and the Warren Spring Laboratory standard
procedure for BWI by ARC.
• Mintek used their batch mill (which conforms to the standard Bond batch ball) to
determine the Bond ball mill Work Index
• The DML’s Magotteaux pilot mill demonstrated that the Unki ore was amenable to
autogenous primary milling and there was no critical size build-up. A target grind of 35
to 40 per cent -75 µm was achieved at a coarse to fine ore ratio of 25:75
The results for the ARC test work are summarized in Table I:
Table I- Unki BWI test work results benchmarked against Great Dyke ore
Limiting size (μm)
Sample 300 150 106
Unki ore BWI (kWh/t) 11.6 17.7 17.6
Other G Dyke tests: BWI (kWh/t) 15.1 16.5* 18.1
* calculated
The Southern African Institute of Mining and Metallurgy
Platinum 2012
365
Mintek results are shown in Table II:
Table II-Mintek BWI test work results for Unki Ore
Sample Limiting screen
(µm)
F80
(µm)
P80
(µm)
Work Index
KWh/t
Unki ore 75 1906.18 57.56 18.2
150 1906.18 114.02 15.0
The above confirms that the design BWI for the primary ball mill is between 15 kWh/t and 18.2
kWh/t. As will be shown later, for sizing the primary mill motor the Bond method is used. The
design Bond Work Indices for the primary and secondary ball mills are 16.5 kWh/t and 20.5
kWh/t respectively. Inefficiency factors to mimic industrial applications were considered.
Flotation
AARL, ARC, and Mintek conducted laboratory flotation test work on core and bulk samples to
determine the metallurgical response of the Unki ore and develop a flow sheet. The test work
also compared the performances of MF1 and MF2 circuit configurations. The challenge the Unki
ore posed was the presence of a footwall shear containing high levels of altered silicates, raised
levels of talc, and swelling clays. The impact of treating ore that included the footwall material
was also investigated and quantified in the DML pilot plant. The recoveries obtained for the
MF1 and MF2 circuits from different test work campaigns are shown in Table 3.
Table III- MF1 and MF2 recovery
Test work
4E Recovery (%)
MF1 MF2
Mintek pilot 84.6 87.0
DML pilot 77.9 80.4
EMC (simulation) 78.3 84.0
DML’s pilot trials indicated that the 4E PGM recovery dropped by as much as 7.7 per cent and
3.7 per cent due to the higher hydrophobic, talcose component of the shear zone foot wall
material. This and the above results favour an MF2 circuit configuration.
Design
The Unki concentrator MF2 flow sheet is shown in Figure 2. The flow sheet included an in-
circuit crushing (ICC) or pebble-milling circuit to mitigate potential comminution problems. The
ICC circuit was planned to consist of pebble transfer system to a cone crusher or HPGR. A
pebble stockpile and reload facility would be allowed to provide optimum mill load control.
T
T
T
The
Plat
36
Ho
ea
A s
rec
ph
ter
po
pri
mi
me
Th
Un
de
m2
It s
go
Th
in
Sou
tinum
6
owe
sy
stra
cen
ase
rm
we
ima
lls
edi
e c
nive
sig2) w
sho
od
e o
Fig
uthe
m 20
eve
fut
ate
ntly
e F
ex
er r
ary
to
a.
com
ers
gn b
wer
oul
di
ove
gur
ern A
012
er,
tur
egic
y i
FAG
xpa
req
y a
o ca
mm
sity
by
re s
ld
sch
era
e 4
Afric
th
re r
c d
nst
G o
ans
qui
nd
ate
min
y o
JK
spe
als
har
all f
4).
can I
his
ret
ec
tall
ope
sion
rem
d se
er f
nut
f C
MR
eci
so
rge
flow
Ex
nsti
cir
rof
isio
led
era
n p
me
eco
for
tion
Cap
RC,
fie
be
e ch
w s
pa
tute
rcu
fit w
on
d w
atio
pla
ent
ond
r fu
n c
pe
, 14
d f
no
har
she
nsi
e of
it w
wa
wa
with
on
ns
ts.
dar
utu
circ
To
4 m
for
ote
rac
eet
ion
Min
wa
as e
as
hin
if
. B
Est
ry
ure
cui
own
mm
th
ed
cter
t of
n fo
ning
s d
ena
ma
n th
ore
Bon
tim
mi
ex
t d
n,
m sl
e d
th
rist
ffe
oot
and
def
abl
ade
he
e c
nd
mat
lls
xpa
des
co
lot
disc
at
tics
red
tpri
d Me
fer
ed
e to
gr
cha
an
ted
re
ans
sign
nfi
wi
cha
Gr
s, e
d f
int
Fig
etallu
red
in
o st
rou
ara
d S
gr
esp
sion
n w
rm
idt
arg
rea
eve
lex
al
gur
urgy
d d
th
tan
up;
cte
SM
ros
pect
n p
wa
ming
h g
ge g
at D
en w
xibi
low
re 2
y
dur
he b
nda
th
eris
MC/
ss p
tiv
pot
as r
g t
gra
gra
Dyk
wit
ility
wan
2- D
ring
bui
ard
hes
stic
/JK
pow
ely
ten
rev
the
ates
ate
ke
th t
y fo
nce
Des
g th
ilt c
dize
se
cs p
m
we
y. H
ntia
view
e e
s a
s.
or
the
or
es w
sign
he
circ
e th
co
pe
mod
er r
How
al a
we
ffic
nd
res
e ra
fut
we
n fl
pr
cui
he
nse
rm
dels
req
we
and
ed
cac
d an
ge
aise
tur
ere
low
roje
t.
siz
erv
mitt
s w
quir
eve
d t
by
cy
n o
ene
ed
re e
m
w sh
ect
zes
vat
ed
wer
rem
er,
the
y th
of
ope
era
hy
exp
ad
hee
t ri
of
ive
, a
re
me
5.2
e fu
he
th
en a
ally
ydr
pan
e.
et f
sk
f th
ely
and
us
nts
2 M
utu
Ce
he
are
fo
rop
nsio
Th
for
re
he U
siz
d h
ed
s w
MW
ure
ent
de
ea
orm
pho
on
e f
Un
evie
Un
zed
en
to
wer
W m
po
tre
sig
of
m s
obic
(s
futu
nki (
ew
ki
d u
ce
o e
re
mo
ote
e fo
gn.
7.5
slur
c g
ee
ure
(3)
; b
mi
unit
fit
esti
2.7
oto
ent
or
A
5 p
rrie
an
th
e IC
but
lls
ts
tte
ma
7 M
ors
tial
M
Afte
per
es
gu
he a
CC
im
wi
co
d w
ate
MW
we
us
ine
er
ce
wit
e c
aer
is s
mpo
th
uld
we
e fi
W a
ere
sag
era
a
ent
th
con
rial
sho
ort
mi
d t
ell w
na
nd
e s
ge
als
rev
(e
low
nte
l ph
ow
an
ills
hu
wit
l in
d 3
sele
of
Re
vie
qu
w v
nt.
hot
wn i
tly
th
us p
th
nst
.7
ect
st
ese
w
iva
vis
.
to
n F
y a
hat
per
Un
tall
MW
ted
ee
earc
of
ale
co
of
Figu
re
ha
rm
nki’
ed
W
fo
l g
ch
th
nt
siti
th
ure
lat
ad b
it
’s l
m
for
or
grin
at
he
to
ies
e p
e 3
tive
bee
firs
lon
mot
r th
bo
ndi
t th
lin
2.0
s an
pla
.
ely
en
st-
ng-
tor
he
oth
ng
he
ner
05
nd
nt
F
F
igu
Figu
ure
ure
4-U
3-
Unk
Fle
ki c
exib
con
ble
nce
de
ntr
esig
rato
gn f
or a
for
aer
the
rial
e U
l vie
T
Unk
ew
The S
ki IC
w (M
Sout
CC c
Mar
ther
circ
rch
rn Af
cuit
20
frica
t
011
an In
1)
nstittute of MMining aand
Plat
Met
tinum
tallu
m 20
36
urgy
012
67
The Southern African Institute of Mining and Metallurgy
Platinum 2012
368
Process description
The Unki concentrator design is for a nominal capacity of 120 kt/month using a 20 foot
diameter by 28 foot length primary mill with 5.2 MW installed mill motor power as a ROM AG
mill in closed circuit with a classification screen (with 0.63 mm to 0.8 mm aperture opening)
and a 20 foot diameter by 28 foot length secondary mill with 5.2 MW installed mill motor
power as a ball mill in closed circuit with a classification cyclone cluster for secondary regrind.
Mills were sized to enable a future twin-module primary milling configuration – future
alternative milling equipment could then be fully evaluated.
The flotation circuit consists of a short residence-time, primary roughing stage comprising
3x70m3 flotation cells and primary cleaning and re-cleaning stages. Primary rougher tails report
to the secondary mill ahead of scavenger flotation, and cleaning and re-cleaning. Two
concentrates are produced, a high-grade concentrate from the primary re-cleaner flotation
circuit and a medium-grade concentrate from the secondary re-cleaner circuit. These combine
to form the final concentrate. Open- or closed-circuit cleaning in the scavenger circuit is an
option
Concentrate handling consists of a Magra thickener and a conventional clarifier, and Larox high-
pressure filtration of the thickened concentrate. Tailings thickening circuitry includes guard
cyclones, a high-rate thickener, and clarifier. Final tails are pumped to a tailings storage facility
about 5 km away from the plant.
The original design concept offers great flexibility and expansion potential for Unki. The layout
provided for a future ICC to allow a sequential expansion of a single module, from 120 to 140
kt/month using suitably oversized 20 foot diameter mills and then by utilizing ICC technology to
potentially achieve a throughput of as much as 180 kt/month, once FAG operation had been
proven. This is achieved through an F80 of 300 mm and P80 of 250 µm in the primary AG mill,
with a circuit grind of 36 per cent and 80 per cent -75 µm in the primary and secondary mill
respectively. It is worth noting that mitigation in this design is by:
1. Oversized 20 foot diameter mills
2. Mill and mill drive designed for 30 per cent steel loading, allowing for a SAG and high-
steel ROM ball milling option. The mill was designed for 75 per cent critical speed and 10
per cent circulating load
3. ICC circuit addition in the event of critical size build-up and retention of FAG. The circuit
will be designed for 20 to 30 per cent pebble generation. The primary mill grate
discharge grate will be modified to include ±65 mm pebble ports (design to be
confirmed from operational data)
4. Possibility of converting the mill drive to VSD as mitigation for throughput and FAG
constraints due to unsuitable ore fragmentation.
The Southern African Institute of Mining and Metallurgy
Platinum 2012
369
Ore characterization
Competent ore and a sufficient coarse fraction component are prerequisites for successful
autogenous milling. An extensive review of literature of all previous and current test work and
operations treating similar ore from mechanized and non-mechanized mined operations was
conducted. Ore hardness and grindability tests demonstrated that the ore could be classified as
hard. The results are shown in Table IV. These results were also used to estimate power
requirements for comminution circuit design.
Table IV- Summary of comparative ore hardness test and operational results
SAMPLE DWT BWI (kWh/t)
A b A*b Ta 106 µm 75 µm
2004 Test
work 67.6 0.62 41.9 0.38 19.2 20.5
May 2011 78.5 0.67 52.6 0.37 19.7 21.1
Analysis of t10 (size passing 10 per cent of the original particle size) and unit specific energy
(Ecs) established that the Unki ore displays the normal trend of decreasing slope with
decreasing energy (Ecs values). A study of the distribution of the relative density of particles
showed no evidence of bimodality (that is, no evidence of dense particles that could build up in
the mill and cause mill load problems).
Two measures were put in place to cater for possible risk of failing to achieve the target primary
grind. An in-circuit crusher circuit was provided for in the layout, but its installation was
deferred. The primary mill liners were designed for both FAG and SAG operation; the first set
was designed conservatively, as is normal practice.
TThe
Plat
37
Pil
bu
the
Fee
Ab
lite
rep
An
sim
co
>1
It i
it t
Sou
tinum
0
ot
lk
ere
ed
bou
era
pre
nglo
mila
nte
90
is i
ten
uthe
m 20
pla
ore
e w
or
ut
atu
ese
o A
ar
ent
mm
mp
nds
ern A
012
ant
e s
was
re s
65
re4
enta
Am
dim
ts o
m).
por
s to
Afric
t te
sam
no
size
p4,5
avi
meri
me
of
. Th
rta
o bu
can I
est
mp
o e
e
per
ind
ity
ica
ens
coa
he
nt
uild
nsti
t w
le
vid
ce
dic
of
n
ion
ars
Un
tho
d u
tute
wor
co
den
ent
cate
th
Pla
ns
ser
nki
ou
up a
e of
k c
nfi
nce
t o
ed
he b
atin
ha
fr
RO
gh
and
Min
con
rm
e of
of
th
bul
num
d d
act
OM
to
d lo
ning
F
ndu
med
f cr
the
hat
lk o
m
dem
tio
M o
o m
oad
and
Figu
uct
d th
riti
e
t t
ore
Pro
mo
ns
ore
mini
d t
d Me
ure
ed
hat
cal
Ro
he
e sa
oce
ons
(th
siz
im
he
etallu
5-
at
t th
l si
M
m
am
ess
stra
his
ze
ize
m
urgy
BW
t A
he
ze
o
max
mple
s G
ate
pa
dis
e th
ill.
y
WI v
Ang
Un
bu
re
xim
e re
Gro
ed t
art
stri
he
vs.
glo
nki
uild
sa
mum
ep
up
tha
icu
ibu
am
lim
Am
or
d-u
am
m
res
o
at
ular
utio
mou
miti
me
re
p in
ple
co
sen
n a
FA
r m
on
unt
ing
ric
wa
n t
e w
ars
nte
a r
G
mill
cu
t of
g siz
an
as
he
was
se
d a
rec
pri
ha
rve
f v
ze t
Pl
am
m
s i
fra
a m
en
ma
ad
e in
ery
tes
ati
men
ill.
in
act
maj
tly
ary
be
ndi
y fi
st w
inu
nab
th
tion
or
co
y m
een
ica
ine
wor
um’
ble
e
n s
po
om
mill
n d
ted
e m
rk r
’s D
to
co
sho
ote
mm
ing
esi
d 3
mat
resu
DM
o a
ars
oul
enti
iss
g co
ign
30 p
eri
ults
ML
uto
se
ld
ial
ion
ou
ned
pe
al
s
fac
oge
fra
be
ris
ned
ld
d fo
r c
(sa
cilit
ene
act
e ±
sk.
d F
be
or a
en
ay <
ty
eou
tion
±15
Th
FAG
e su
a R
t >
<1
in
us
n (
5-2
e e
G p
ucc
ROM
>20
m
Ru
mi
(>1
5
exp
prim
ces
M
00 m
m)
uste
illin
106
pe
per
ma
ssfu
ore
mm
) in
enb
ng,
6 m
r c
rien
ry
ul f
e o
m (
n th
bu
, an
mm
cen
nce
m
for
of 2
(Fig
he f
rg
nd
m).
nt.
e o
ill
hi
20-
gur
fee
on
th
Th
Th
f th
wi
igh
-30
re 4
ed
n a
hat
he
he
he
ith
her
0%
4).
as
T
Mi
A c
op
mi
ou
als
tha
an
im
Ke
rat
wid
dis
a g
ind
13
To
an
be
ill l
cau
era
ll l
tm
so
at t
gle
ple
y a
tio
de
sch
gra
dica
m
ca
d o
en
line
utio
atio
ine
mos
an
the
e o
em
asp
(S
r li
harg
ate
ati
m)
ate
opt
ch
ers
ou
on
er
st p
aly
e li
of 7
en
pec
/H
ne
ge
di
on
), a
r fo
ted
han
s
s a
bu
de
par
yse
ne
70o
nted
cts
). A
er t
gr
isc
s t
a di
or t
d fo
nge
app
ut a
sig
rtic
d t
r do m
d.
of
As
top
ate
ha
ha
sch
the
or
ed t
pro
abl
gn
cles
the
des
mea
th
th
p, w
e fo
rge
t t
har
e p
Mo
to
oac
le t
wa
s w
e t
sign
asu
e l
he
whi
or
e o
he
rge
pos
oly
a h
ch w
to
as
wer
raj
n w
ure
ine
mi
ich
op
ope
Un
e gr
sib
-Ch
har
wa
op
pro
re a
ect
was
d f
er d
ll w
ad
tim
eni
nki
rat
bilit
hro
rde
Fi
as a
per
opo
ana
tor
s su
fro
des
wa
dds
mum
ng
or
te o
ty o
om
er m
igu
ado
ate
ose
aly
ries
uita
m
sig
s u
s to
m
of
re
ope
of S
e (
mat
re
opt
e in
ed.
yse
s o
abl
th
n w
uni
o l
mi
f b
cou
eni
SA
(ap
ter
6-U
ted
n S
Th
d u
of t
le f
e b
we
dir
ine
ll p
betw
uld
ing
G o
ppr
rial
Unk
d t
SAG
his
usi
the
for
bas
re
rec
er l
per
we
d p
g of
ope
oxi
.
ki t
o m
G m
s de
ng
e o
r AG
se,
the
ctio
life
rfo
een
oss
f 14
era
ima
ria
mit
mod
esi
M
out
G/S
an
e f
ona
e. D
rm
n 1
sib
4 m
atio
ate
l m
tig
de.
gn
MillT
mo
SAG
nd
ace
al t
Due
man
12
ly
mm
on
ely
mini
ate
. W
w
Tra
ost
G o
85
e a
the
e c
nce
an
ge
m w
the
2
ing
e p
Wor
was
aj s
pa
ope
5o f
and
e tr
on
(o
d
ne
was
e d
pe
g RO
pot
rkin
re
soft
art
era
for
d tr
rail
sid
ope
16
rat
sp
des
r c
OM
ten
ng
evie
tw
ticl
atio
r th
rail
ing
der
en a
m
te a
pec
sign
cen
M si
ntia
wi
ew
are
es.
on.
he
ing
g a
rati
are
mm
a c
cifie
n te
t c
T
ze
al f
th
wed
e (
. In
Th
tra
g a
ang
ion
ea)
is
crit
ed.
eam
chr
The S
dis
flaw
th
d b
Fig
n b
he
aili
ng
gle
n w
. T
re
ica
.
m s
om
Sout
stri
ws.
e m
y J
gur
bot
op
ng
les
wa
was
Typ
eco
al s
shi
miu
ther
but
. Li
mil
JKM
e 7
h c
ptim
an
s an
as
al
ica
om
ize
ed
m)
rn Af
tio
ine
ll v
MR
7).
cas
mu
ngl
nd
ma
so
ally
me
e (e
aw
). T
frica
n
ers
ven
C6.
A
ses
m
e)
th
ade
giv
y fo
end
esti
way
The
an In
w
do
. T
th
s th
lin
wa
e l
e s
ven
or m
ded
ima
y f
e lin
nstit
ere
or a
he
ird
he
ner
as
ine
ste
n t
mil
d.
ate
rom
ner
tute
e d
and
tr
d-pa
fin
de
rec
er s
ep
to t
ls i
As
ed
m a
r m
of M
des
d li
raje
ary
ndi
esig
com
spa
er,
the
in t
th
at
a h
mat
Mini
sign
ne
ect
yt l
ng
gn
mm
aci
, re
e d
thi
her
ap
hard
ter
ng a
ned
r s
tori
line
s c
(w
me
ng
esu
esi
s o
re
ppro
de
ial
and
Plat
d f
sup
ies
er
con
with
nd
to
ulti
ign
ope
ha
oxi
r m
ha
Met
tinum
for
ppli
of
ve
nfir
h a
ed
o he
ng
n of
era
d b
ima
mat
as s
tallu
m 20
37
FA
ier,
f th
nd
rme
fa
an
eig
in
f th
tio
bee
ate
ter
sin
urgy
012
71
AG
, a
he
or
ed
ce
nd
ght
n a
he
on,
en
ely
ial
ce
The Southern African Institute of Mining and Metallurgy
Platinum 2012
372
Figure 7- Example of projected trajectories for the Unki primary mill
Operation and post-commissioning
Mill control
The application of Anglo American Platinum’s APC (Advanced Process Control System)
guaranteed the performance of the mill. The installation of a mill feed silo with enough buffer
capacity (>20 hours) ahead of the mill also enhanced the control solution.
Plant performance
The Unki concentrator has been able to operate at above design throughput (approximately
130 kt/month versus design of 120 kt/month) with a primary grind meeting design specification
(approximately 42 per cent -75 µm versus design of 35 to 40 per cent -75 µm) and final grind of
approximately 80 per cent -75 µm. The oversize (>19 mm) material generated from the primary
mill is much less than predicted. Figure 8 is a typical size distribution curve for the Unki
concentrator comminution circuit.
���
���
��
���
��
��
��
�� �
���
���
���
��
��
��
���
���
��
��
���
���
�����
��
��
����
���
���
��
��
�
���
����
���
���
���
���
����
���
�� �
��
��
���
��
���
�
��������
���
��
��
���
����
���
��
���
���
����
��
����
���
����
�
��
��
��
��
��
��
�
�
��
���
��
�
����
��
���
����
�� �
�
��
��
��
�� �
���
�
� ��
���
����
����
����
����
����
����
����
����
����
����
����
��
���
���
� ��
����
�� �
���
��
���
��
��
���
��
��
���
�� �
�
�
�
�
�
�
�
�
�
�
�
�
���
�
��
���
���
���
���
���
�
���
���
��
���
���
��
��
�
�
�
��
��
���
�
� ��
!"#
�� �
��
��
���
����
����
��
����
���
���
�
� �
�
���
���
$%
� �
# �
���
����
�
��
��
�
��
��
��
���
��
���
� �
� &�
% '
����
���
���
��
��
���
��
�
��
� �'
��
��
���
��
���
'��
��'�
����
���
�
��
���
� �
���
����
' ��
��
� �
���
����
�
��
��
��
����
�
���
��
���
���
���
��
����
�
���
����
����
��
�� �
����
��
���
���
����
����
���
� �
� �
� �
���
�
��
�����
�����
���
���
��(
��
��
���
���
���
� �
���
���
��
��
��
���
���
���
� �
�
( �)
��
���
����
��
���
�� �
��
��
'�
���
�
�
��
���
�)
�'
)��
���
���
��
��
� �
�)�
� �
&$
���
���
���
�
�� ��
�� �
���
��
���
��
���
���
���
���
����
��
��
�
���
�
��
���
����
���
����
����
� ��
��
����
����
��
���
���
%� '
��
� �
��
��
��
�����
�����
���
���
�� �
��
��
��
���
���
���
�*�
'��
��
���
�
����
�
���
���
���
���
��
��
�$�
���
��
���
� �
��
� �
���
���
�
��
��
� �
���
��
��'
���
�
��
���
���
��
���
��
��
��
���
��
���
����
��
����
��
���
����
��
� �
����
��
���
�
����
��
��
����
�+
�'��
���
&���
����
����
���
��
��
��
��
���
���
��
����
�
� �
+�,
� ��
��
��
��
��
� �
� ��
�'
�
� '�
�
����
��
���
�����
���
, $%
���
��
�
��
����
�� �
��
' �
��
���
��
�
��
���
���
����
%
���
��
���
���
���
��
��
���
��
���
��
�
�
��
����
�� �
���
���
��
���
���
���
���
���
� �
���
� �
�
�
�����
�����
��
� ��
' -
���
���
����
����
���
���
��
� �
�� �
��
���
����
����
���
��
-.��
��
� /
�
���
���
��
���
���
���
�
���
���
�
��
� $
��
��
�
���
��
��
���
�
���
���
��
���
�����
���
���
$%
���
�� �
����
�����
� �
���
���
���
� �
�
��
���
�����
�� �
� �
�!
���
���
���
�����
���
�
�"� �
���
����
��
�
��
����
���
��
� �
���
���
����
��
����
���
��
���
���
��
����
��
���
��
���
���
���
���
��
� �
���
���
�
���
��
���
-.
���
���
���
��
���
� �
���
���
��
��
� �
���
.�0
��
���
���
'�
���
����
���
���
���
����
��
���
#���
� �
"�
�
���
����
����
�"�
���
���
���
����
1�
��
���
%�
���
��
� ��
2 ��
��
��
���
�
�����
����
�1
��
���
�3�
��
�
��
4�
� �
���
��
��
�
TThe
Plat
37
Ho
op
wh
hig
silo
occ
co
co
Sou
tinum
4
owe
era
hich
gh
o l
cas
nsi
nsi
uthe
m 20
eve
ato
h w
silo
ev
sio
ide
ide
ern A
012
er,
or m
wo
o le
els
na
ered
ered
Afric
th
ma
rse
eve
s, d
lly)
d
d e
can I
he
ana
ens
els
diff
) th
du
exc
nsti
in
age
s w
. T
fer
he
rin
cess
tute
tro
es. A
whe
he
en
op
ng
siv
e of
odu
As
en t
m
tia
per
de
e.
Min
uct
ca
the
ill f
ally
rato
esig
Fig
Fi
ning
tion
an b
e m
fee
b
or
gn,
gur
gur
and
n o
be
mill
ed
etw
dr
b
e 1
re 1
d Me
Fig
of
im
l fe
silo
we
aw
ut
10
10-
etallu
gur
in
mag
eed
o h
en
ws f
as
sho
-Or
urgy
re 9
suf
gine
d si
has
th
fro
s t
ow
re o
y
9- P
ffic
ed,
ilos
s tw
he
om
the
ws e
on m
Pow
cien
, by
s a
wo
ch
th
e b
exa
mil
wer
nt
y it
re
dis
hut
e c
brie
am
ll fe
r dr
co
ts v
ru
sch
tes
chu
ef
ple
eed
raw
oar
ver
unn
har
. W
ute
wa
es o
d co
w p
rse
ry n
ning
rge
Wh
es p
as
of t
onv
lot
fr
nat
g lo
e ch
hen
pro
for
the
vey
fo
rac
tur
ow
hut
n s
opo
r a
e fe
yor
r th
tio
e t
w. T
tes
uch
ort
a p
eed
(he
he
on
the
The
s an
h s
ion
pla
d o
elm
mi
cre
e RO
e st
nd
situ
nat
nt
re
met
lls
eat
OM
tra
the
uat
tely
fit
on
t as
tes
M o
teg
e o
tion
y. A
t f
n th
s sc
s d
ore
gy
ore
ns
A c
or
he f
cale
dist
is
is
e w
oc
coa
pu
fee
e)
tur
pr
the
wou
ccu
arse
urp
ed
ba
on
ere
uld
ur
e/f
pos
co
nc
e t
efo
se
(th
fine
se
nve
es,
to s
re
gre
hes
e o
th
eyo
, w
seg
to
ega
eo
ore
is
or.
whi
gre
m
ate
ccu
sp
co
ich
ega
main
e at
ur
plit
ost
th
tio
nta
t lo
on
t w
w
he
on,
ain
ow
nly
was
was
The Southern African Institute of Mining and Metallurgy
Platinum 2012
375
Primary mill liner wear
The first set of liners were of an abrasion-resistant, air-hardened pearlitic chrome-molybdenum
alloy steel approximately 2% Cr). These liners were deliberately selected for possibility of
adding steel (SAG operation) due to ’critical size potential’ and were expected to last between 9
and 12 months. The liners were replaced after nine months. After demonstrating that FAG
primary milling is feasible, the liner material has been changed to a martensic abrasion-
resistant high-chrome iron (16.5-24.5% Cr). This material is not tolerant to steel media. The
high-chrome iron liners are approximately 32 per cent more expensive than the Moly-Cr liners,
but last approximately twice as long. This has been confirmed to date by monitoring of liner
wear. Table V demonstrates the benefit of using a more durable and expensive liner for the
primary mill operated in FAG mode. The throughput for the new liners is a projected from wear
rates over a six month period.
Table V-New and old mill liner comparison
Liner type Throughput to
replacement (t)
Liner cost (R) Liner cost
(R/t)
Liner life
(months)
Old (Moly-Cr : Low
Cr)
725 012 6 392 585 8.82 9
New (High Cr)* 2 625 000 10 010 291 3.81 21
Estimated from liner wear monitoring on plant
Platinum recovery
The current ROM material contains significant footwall shear zone material, which is depressing
recovery due to raised hydrophobic gangue levels and hence greater competition with slower-
floating value minerals. Target recoveries are for mechanized mining cut excluding the FWSZ.
Figure 11 represents a cross-section through the typical Unki mineral width (mining cut of 2.2
m), with platinum grades peaking at the Main Sulphide Zone (MSZ) and decreasing into the
hangingwall and footwall. The zone in the footwall is characterized by presence of hydrophobic
altered silicate talcose material that impacts negatively on plant PGM and BMS recoveries.
T
T
The
Plat
37
Th
pla
sho
20
82
80
fac
to
ev
FW
Sou
tinum
6
e s
atin
ow
11
pe
pe
cto
M
ide
WSZ
uthe
m 20
suc
num
wing
to
er c
er
ors
arc
enc
Z.
ern A
012
cce
m
g a
o M
cen
ce
ca
ch
ce o
Afric
ssf
rec
a po
Mar
nt,
nt.
use
20
of
can I
ful
cov
osi
rch
an
In
ed
012
a s
nsti
op
ver
tiv
20
nd
n th
by
2, t
stea
tute
per
ry.
ve t
012
wit
he
y re
he
ady
e of
ati
By
tre
2.T
thi
Au
edu
pl
y u
Min
F
on
co
nd
he
in 7
ugu
uce
ati
upw
ning
Figu
of
ont
sin
av
7 m
ust
ed t
nu
war
and
ure
f th
tinu
nce
ver
mo
20
ton
um
rd
d Me
e 11
he
uou
e c
rag
nth
011
nna
re
tre
etallu
1-C
FA
usl
com
e t
hs
1, t
age
cov
end
urgy
Cros
G m
y a
mm
targ
of
the
e, m
ver
d to
y
ss-s
mil
att
miss
get
op
e s
mil
ry
ow
sec
ll in
ain
sion
t p
per
pik
ll re
ha
ward
ctio
n th
ning
nin
lat
rati
ke
elin
s b
ds
on t
he
g t
ng.
inu
ion
of
nin
bee
the
thr
Un
the
Fig
um
n (J
up
ng a
en
e t
oug
nki
e ta
gur
m re
un
p to
and
co
arg
gh
co
arg
re
eco
ne 2
o 8
d a
nsi
get
Un
om
et
12
ove
201
86
a hi
iste
ted
nki
mu
gr
sh
ery
11)
pe
igh
ent
d 82
mi
uni
ind
how
ov
) th
er c
h he
t, a
2 p
nin
tio
d, t
ws
ver
he
cen
ead
ave
per
ng w
on c
the
the
th
pla
nt w
d g
erag
r ce
wid
circ
e p
e U
e p
atin
wa
gra
gin
ent
dth
cui
lat
Unk
pro
nu
s d
de
ng a
t, d
h
it c
inu
ki r
ojec
m
due
. F
aro
des
can
um
rec
ct l
rec
e t
Fro
oun
pit
n al
re
cov
life
cov
o m
om
nd
te t
so
eco
ery
e fo
ver
me
Se
78
the
be
ove
y fr
or t
ry h
etal
ept
8 pe
e im
e se
ery
rom
this
had
l ac
em
er
mp
een
ha
m J
s c
d r
cco
mbe
ce
act
n in
as b
Jan
irc
ea
oun
er 2
nt
t o
n th
bee
nua
uit
che
nti
20
wi
f th
he
en
ary
t is
ed
ng
11
ith
he
Pro
Fig
mo
be
odu
gur
ont
en
uct
re 1
ths
ab
tion
13
s of
ble
n r
sho
f co
to
ram
ow
om
o m
mp-
ws t
mm
main
-up
the
iss
nta
p
e U
sion
ain
nk
nin
th
ki co
ng t
e h
on
the
high
cen
e p
h t
Fig
ntr
lan
hro
gur
rato
nt w
oug
re 1
or
wa
gh
12-U
pro
s a
put
Un
od
abl
ts c
ki p
uct
e o
con
pla
tio
of e
nsi
tin
n r
exc
ste
um
ram
cee
ent
m re
mp-
edi
tly.
T
eco
-up
ng
The S
ove
p fr
de
Sout
ery
rom
esig
ther
m s
gn
rn Af
sta
na
frica
rt o
am
an In
of
ep
nstit
pro
lat
tute
odu
te c
of M
uct
cap
Mini
tion
pac
ng a
n. W
city
and
Plat
Wi
y, a
Met
tinum
thi
and
tallu
m 20
37
in 1
d h
urgy
012
77
10
as
TThe
Plat
37
Co
Fig
gri
ran
Sou
tinum
8
omm
gur
nd
nge
uthe
m 20
min
re 1
d a
es.
ern A
012
nut
14
s p
Afric
tio
sh
pro
can I
on c
how
odu
nsti
circ
ws
uct
tute
cui
th
ion
e of
Fig
it p
e g
n i
Min
gur
perf
grin
ncr
ning
re 1
for
nd
rea
and
13-
rma
ac
ase
d Me
-Un
an
chi
ed
etallu
nki
ce:
eve
ab
urgy
pro
: gr
ed
bov
y
odu
rin
fro
ve
uct
nd d
om
na
ion
dat
m c
me
n (k
ta
com
epl
kt/m
mm
ate
mo
miss
e c
onth
sio
cap
h) f
nin
pac
fro
ng.
city
m c
Th
y, b
com
he
but
mm
da
t st
miss
ata
till
sio
sh
cl
nin
how
os
ng
ws
e t
sli
to
igh
th
ht d
e t
dec
tar
cre
rge
ease
et g
es
grin
in
nd
Co
Fu
Un
me
op
No
ah
fra
A
sig
thi
int
onc
lly
nki
eas
era
•
•
•
o as
ea
agm
les
gnif
is a
terr
lus
au
M
sur
ati
sse
d o
men
sso
fica
app
rog
sio
uto
ine
res
ng
R3
req
R4
RO
R5
est
ess
of t
nta
on
ant
pro
gat
n
oge
es C
pr
co
.81
qui
.80
OM
.80
tim
me
the
atio
lea
tly
oac
tion
no
Con
rov
osts
1 –
ire
0 –
pr
0 –
mat
ent
e p
on
arn
to
ch
n o
ous
nce
ide
s, a
–
me
– sa
rim
– sa
ed
t h
prim
in
nt
sh
mu
of t
pr
ent
ed
am
red
ent
ave
mary
ave
po
as
ma
co
fro
har
ust
ech
rim
tra
for
ou
duc
ts
ed
y b
ed m
ow
be
ry
nju
om
reh
t b
hn
mar
tor
r in
nti
ced
gr
ball
mil
wer
een
mi
unc
th
hold
e s
ica
ry m
r.
n th
ing
d
ind
l m
l p
req
n m
ill.
ctio
he
der
sup
al d
mil
It h
he
g to
pri
din
mill
pow
qu
mad
On
on
U
r v
ppo
ata
llin
has
de
o ap
ima
g m
wer
ire
de
ngo
wi
nk
alu
ort
a a
ng w
s n
esig
pp
ary
me
r. C
eme
of
oin
th
i e
ue
ed
va
F
wa
ot
gn.
rox
y m
edia
Com
ent
th
ng o
th
exp
by
by
ilab
Figu
s s
be
Th
xim
mil
a f
mp
ts f
e r
opt
e m
per
y fu
y a
ble
ure
suc
een
he b
mat
ll
from
ari
for
red
tim
min
ien
ully
a t
e.
14
cce
n n
be
tely
lin
m
ng
th
duc
miza
nin
nce
y ev
ho
4-U
ssf
ece
nef
y R
er
est
g ac
he p
ctio
ati
g t
e is
val
rou
nki
full
ess
fits
14
co
tim
ctu
pri
on
on
ea
s t
lua
ugh
i gr
ly i
sar
s a
4 pe
ost
mat
al
ma
in
in
m.
tha
atin
h u
rind
imp
y t
re
ert
ts,
ted
po
ary
ca
nitia
at
ng
und
d d
ple
to i
sig
on
ex
d st
owe
m
pit
ativ
de
the
der
T
ata
em
mp
gnif
m
xcl
tee
er c
ill o
tal
ves
sig
e p
rsta
The S
a
ent
ple
fica
ille
ud
el c
con
op
as
s o
gne
pot
and
Sout
ted
eme
ant
ed
ing
con
nsu
era
a
on s
ers
ten
din
ther
d a
ent
t as
an
g s
nsu
um
ate
res
site
of
ntia
ng
rn Af
at A
t th
s re
d m
sav
um
pti
ed a
sul
e i
f p
al d
of
frica
Ang
he
efle
ma
vin
pti
ion
as
t o
ncl
pro
des
th
an In
glo
co
ect
de
gs
on
n on
a R
of n
lud
oce
sign
he
nstit
o A
omm
ted
up
in
n fo
n t
ROM
not
de
ssi
n e
ore
tute
Ame
min
d in
p a
n r
or
he
M
t in
op
ng
env
ebo
of M
eric
nut
n th
s fo
red
mi
pr
pri
nclu
tim
p
velo
ody
Mini
can
tio
he
ollo
duc
ll o
rim
ima
ud
miz
lan
ope
y a
ng a
n P
on m
red
ow
ced
ope
mar
ary
ing
ati
nts
e.
and
and
Plat
Plat
mit
duc
ws:
d r
era
y m
y ba
g cr
on
ca
Ho
d s
Met
tinum
tin
tiga
ctio
rel
atin
mil
all
rus
n of
an
owe
en
tallu
m 20
37
um
ati
on
ini
ng
l an
mi
shi
f o
ad
eve
sib
urgy
012
79
m’s
ng
in
ng
as
nd
ill.
ng
ore
dd
er,
ble
The Southern African Institute of Mining and Metallurgy
Platinum 2012
380
Acknowledgements
Anglo American Platinum for permission to publish the paper.
My co-authors Chris Rule and Sarah Makumbe.
Unki Platinum Mines concentrator team for living the FAG primary mill dream and making it a
success story.
References
1. Hey. P. V. Platinum Group Mineral Characterisation of Various Channel Samples from the
Unki Deposit, Great Dyke Mineralogy. Report 02/116. Anglo Research Centre, Germiston,
South Africa, 2003.
2. Schouwstra, R. P. Mineralogical Characteristics of PGE Deposits of the Great Dyke
Zimbabwe (Literature Research). Anglo Research Centre, Germiston, South Africa, 2003.
3. Nyakudarika, A. Unki Mine - Process Design Criteria 120 ktpm Concentrator Plant Design -
RoM AG Milling Option. Anglo American Platinum - Process Technolgy Division, Nov.
2007.
4. Sweet, J. and Powell, M.S. Assessment of design of Unki Comminution Circuit. Julius
Kruttschnitt Minerals Research Centre (JKMRC), University of Queensland, Australia,
August 2007.
5. Napier-Munn, T.J., Morrell, S., Morrison, R.D., and Kojovic, T.. Mineral Comminution
Circuits : Their Operation and Optimisation. Julius Kruttschnitt Minerals Research Centre
(JKMRC), University of Queensland, Australia, 1999.
6. Powell, M.S. Unki liner and discharge design review. ulius Kruttschnitt Minerals Research
Centre (JKMRC), University of Queensland, Australia. Report 1. JKTech, August 2007.
The Southern African Institute of Mining and Metallurgy
Platinum 2012
381
The Author
Tony Nyakudarika, Principal Process Engineer, DRA Mineral Projects (Pty) Ltd
Tony Nyakudarika is Principal Process Engineer with DRA Mineral Projects. He has been in this
role since October 2011. His responsibilities are to manage process design and commissioning
of client concentrator projects. Prior to joining DRA, Tony was involved in the mining industry
for over sixteen years as a Divisional Engineer, Project Services Manager, and Lead Process
Engineer for Anglo American Platinum in Zimbabwe and South Africa. Over the last seven years
Tony has been involved in design and commissioning of platinum concentrators for both
greenfield and brownfield projects as a Senior/Lead Process Engineer. Some of the projects
Tony was involved in were Amandelbult UG2 Upgrade, Amandelbult Tailings Disposal Upgrade,
and the Unki Concentrator from PFS to operation.
Tony also spent more than fifteen years working for multinational companies in thecChemicals
and household and personal hygiene goods industries in engineering and operation roles in
Zimbabwe and Tanzania.
The Southern African Institute of Mining and Metallurgy
Platinum 2012
382