3886 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
A computer circuit model was constructed from the
2018 survey data using the Metcom Streamline ™ software.
The ball mill mathematical model consists of the discrete
ACGR points. The cyclone mathematical model uses the
discrete size-by-size recovery to underflow points with and
without bypass. Use of discrete values for both eliminates
model fitting and associated errors. The cyclone feed pump
model was its measured (survey) slurry flow rate and total
dynamic head (calculated from cyclone feed pressure, verti-
cal lift from the pump to the cyclone cluster, and piping
losses). Its operating point on the manufacturer’s pump
performance curve, including impellor speed and pump
motor power, was thus verified. This flow determination
requires a very accurate measurement of the circulating
load ratio during the survey, which is only possible with
accurate survey samples. Combining the calculated mill,
cyclone, and pump models with the circuit feed mass flow
and size distribution, the Streamline ™ program then pro-
vides exact duplication of the survey data.
The new desired cyclone solids and water mass bal-
ance can be calculated from the circuit tonnage and water
addition rate at the same circulating load ratio (McIvor,
1984). The Streamline ™ circuit model is then used to find
the desired cyclone separation performance in terms of d50c
to achieve the desired cyclone overflow sizing P80 at the
specified desired circulating load ratio. This, along with
the desired cyclone feed pressure and the installation lay-
out, provides the total dynamic head, which, together with
cyclone feed flow rate and slurry density, defines the new
desired pump performance.
The cyclones are selected (number in operation and
critical dimensions) for the new performance duty based
on established capacity and separation (d50c) performance
guidelines, like those provided by Plitt (1976). Whether
this can be achieved through dimensional changes to the
existing cyclones, or whether (a cluster of) new cyclones are
needed can then be evaluated. Whether the needed cyclone
feed slurry flow and total dynamic head can be achieved
through a speed change with the existing pump and motor,
or whether a new motor and/or pump are needed, is also
evaluated. Note that the circuit modeling first established
the required processing performances of the pump and cyclones
to achieve the desired grinding circuit performance in
terms of CSE. The equipment’s detailed specifications are
then chosen as a subsequent, separate step to achieve the
required processing performances.
In the above example, a grinding circuit tonnage
increase is accompanied by the improved cyclone water
balance. New cyclone clusters and new cyclone feed pumps
and motors were installed. A verification survey in 2021
measured the circulating load ratio to be 479%, while
cyclone overflow percent solids by weight fell to 24.8%
from 33.6%, and CSE at 75 microns increased to 85.0%
from 76.1%. This was reflected in an overall (SAG-ball
mill) circuit efficiency increase.
Additional examples of increasing CSE through
improved cyclone water balance, increasing circulating
load ratio, and implementation of better separation perfor-
mance cyclones may be found in McIvor (2014), McIvor et
al., (2017), and Bartholomew et al., (2018).
Increasing Grinding Rate
The primary variables affecting a given plant ball (or peb-
ble) mill’s grinding efficiency have been found to be grind-
ing media selection, mill feed water addition rate, and the
design of the mill discharge in terms of how it affects the
volumetric slurry hold-up in the mill.
Early examples of plant testing of alternative media size
or shape are given by McIvor et al, 1991 and McIvor et al,
1994. In these two studies, plant surveys were conducted
before and after a change in media shape in one case, and to
compare parallel balls mills making use of different make-
up ball sizes in the other. The Functional Performance
Equation was used to demonstrate a difference in mill
grinding efficiency was associated with different grinding
media usage by taking into account different ore qualities,
CSEs, and mill powers measured during comparative plant
surveys. The initial ball mill circuit survey at Les Mines
Selbaie produced the following Functional Performance
equation at 106 microns.
33.3 t/h =523 kW × 71.0% × 2.31 g/rev
× 0.0388 (t/kWh) /(g/rev)
A follow-up survey approximately one year after conver-
sion (and time to reach equilibrium) of the ball charge
to a different media produced the following Functional
Performance equation.
32.1 t/h =539 kW × 71.5% × 1.69 g/rev
× 0.0493 (t/kWh) /(g/rev)
The measured increase in MGrEff was approximately 25%.
The parallel, wet, open circuit mills at the Quebec
Cartier pellet plant were surveyed and produced the follow-
ing Functional Performance equations at 45 microns. In
this case batch grindability tests were carried out on circuit
(mill) feeds. For mill D, charged with 60% of 25 mm and
40% of 38 mm grinding balls:
122.5 t/h =3,873 kW × 69.1% × 14.6 g/rev
× 0.00313 (t/kWh) /(g/rev)
A computer circuit model was constructed from the
2018 survey data using the Metcom Streamline ™ software.
The ball mill mathematical model consists of the discrete
ACGR points. The cyclone mathematical model uses the
discrete size-by-size recovery to underflow points with and
without bypass. Use of discrete values for both eliminates
model fitting and associated errors. The cyclone feed pump
model was its measured (survey) slurry flow rate and total
dynamic head (calculated from cyclone feed pressure, verti-
cal lift from the pump to the cyclone cluster, and piping
losses). Its operating point on the manufacturer’s pump
performance curve, including impellor speed and pump
motor power, was thus verified. This flow determination
requires a very accurate measurement of the circulating
load ratio during the survey, which is only possible with
accurate survey samples. Combining the calculated mill,
cyclone, and pump models with the circuit feed mass flow
and size distribution, the Streamline ™ program then pro-
vides exact duplication of the survey data.
The new desired cyclone solids and water mass bal-
ance can be calculated from the circuit tonnage and water
addition rate at the same circulating load ratio (McIvor,
1984). The Streamline ™ circuit model is then used to find
the desired cyclone separation performance in terms of d50c
to achieve the desired cyclone overflow sizing P80 at the
specified desired circulating load ratio. This, along with
the desired cyclone feed pressure and the installation lay-
out, provides the total dynamic head, which, together with
cyclone feed flow rate and slurry density, defines the new
desired pump performance.
The cyclones are selected (number in operation and
critical dimensions) for the new performance duty based
on established capacity and separation (d50c) performance
guidelines, like those provided by Plitt (1976). Whether
this can be achieved through dimensional changes to the
existing cyclones, or whether (a cluster of) new cyclones are
needed can then be evaluated. Whether the needed cyclone
feed slurry flow and total dynamic head can be achieved
through a speed change with the existing pump and motor,
or whether a new motor and/or pump are needed, is also
evaluated. Note that the circuit modeling first established
the required processing performances of the pump and cyclones
to achieve the desired grinding circuit performance in
terms of CSE. The equipment’s detailed specifications are
then chosen as a subsequent, separate step to achieve the
required processing performances.
In the above example, a grinding circuit tonnage
increase is accompanied by the improved cyclone water
balance. New cyclone clusters and new cyclone feed pumps
and motors were installed. A verification survey in 2021
measured the circulating load ratio to be 479%, while
cyclone overflow percent solids by weight fell to 24.8%
from 33.6%, and CSE at 75 microns increased to 85.0%
from 76.1%. This was reflected in an overall (SAG-ball
mill) circuit efficiency increase.
Additional examples of increasing CSE through
improved cyclone water balance, increasing circulating
load ratio, and implementation of better separation perfor-
mance cyclones may be found in McIvor (2014), McIvor et
al., (2017), and Bartholomew et al., (2018).
Increasing Grinding Rate
The primary variables affecting a given plant ball (or peb-
ble) mill’s grinding efficiency have been found to be grind-
ing media selection, mill feed water addition rate, and the
design of the mill discharge in terms of how it affects the
volumetric slurry hold-up in the mill.
Early examples of plant testing of alternative media size
or shape are given by McIvor et al, 1991 and McIvor et al,
1994. In these two studies, plant surveys were conducted
before and after a change in media shape in one case, and to
compare parallel balls mills making use of different make-
up ball sizes in the other. The Functional Performance
Equation was used to demonstrate a difference in mill
grinding efficiency was associated with different grinding
media usage by taking into account different ore qualities,
CSEs, and mill powers measured during comparative plant
surveys. The initial ball mill circuit survey at Les Mines
Selbaie produced the following Functional Performance
equation at 106 microns.
33.3 t/h =523 kW × 71.0% × 2.31 g/rev
× 0.0388 (t/kWh) /(g/rev)
A follow-up survey approximately one year after conver-
sion (and time to reach equilibrium) of the ball charge
to a different media produced the following Functional
Performance equation.
32.1 t/h =539 kW × 71.5% × 1.69 g/rev
× 0.0493 (t/kWh) /(g/rev)
The measured increase in MGrEff was approximately 25%.
The parallel, wet, open circuit mills at the Quebec
Cartier pellet plant were surveyed and produced the follow-
ing Functional Performance equations at 45 microns. In
this case batch grindability tests were carried out on circuit
(mill) feeds. For mill D, charged with 60% of 25 mm and
40% of 38 mm grinding balls:
122.5 t/h =3,873 kW × 69.1% × 14.6 g/rev
× 0.00313 (t/kWh) /(g/rev)