XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3 3885
milling, autogenous and semi-autogenous milling, and
high-pressure grinding (Amestica et al., 1993 Hinde &
Kalala, 2009). Among other uses (like the study of mill
media sizing effects, described later), it has also been used
to model the Bond Work Index test (Ciribeni et al., 2020).
While it relies on the above-noted assumption, it avoids
the bewildering complexity, numerous assumptions, inac-
curacies associated with model fitting, and demonstrated
falsehoods of classical population balance-based computer
modelling and simulation (McIvor et al., 2023).
The same survey data as described above demonstrates
an integral link between the Functional Performance mill
grinding rate (MGrR) and the apparent cumulative mill
grinding rate (ACGR) at the same size of interest. They are,
in fact, different methods of calculating the same grinding
rate parameter, that is, the grinding rate of material desig-
nated as target plus size material per unit of energy applied
to target plus size material in the mill. The Functional
Performance mill grinding rate (a value calculated above
0.404 t/kWh) assumes that the percentage of plus size in the
mill is the average of that size in the mill feed and discharge.
The ACGR calculated at the same size (a value calculated
above 0.402 t/kWh) uses the same mill feed and discharge
size distributions and the same mill power but assumes
plug flow and first-order rate disappearance (i.e., constant
grinding rate per unit of plus size remaining as it dimin-
ishes in quantity). That they differ only to the third decimal
place for typical closed-circuit grinding mills attests to both
their fundamental difference and essential sameness. And
fortunately, used in conjunction with cyclone discrete size
separation performance and pump head and capacity per-
formance, this provides the ability to create a circuit mod-
elling program of high accuracy. Such knowledge of not
only how to model the mill reliably but also pump and
cyclone process application engineering provides a direct
step-by-step path to circuit optimization, as described in
the examples that follow.
EXAMPLES
Increasing Circuit Classification System Efficiency, CSE
The 2018 ball mill circuit survey at Fekola, referenced in
Clark (2023), showed a CSE of 76.1%. This compared to
values of CSE approximately 10 percentage points higher
in the Metcom database in circuits with fully optimized
pump and cyclone equipment. The circulating load ratio
was already suitably high (442%) and modern cyclones
with high sharpness of separation were being utilized. The
remaining opportunity was to improve the cyclone water
balance by increased feed water, subsequent reduction of
bypass of fines being returned to the mill, and therefore
increased CSE, as achieved in multiple cases in the past
(McIvor, 2014 McIvor et al, 2017).
Source :Clark, et al., 2023
Figure 3. Apparent cumulative grinding rate (by size class)—Fekola ball mill, 2018 circuit survey
milling, autogenous and semi-autogenous milling, and
high-pressure grinding (Amestica et al., 1993 Hinde &
Kalala, 2009). Among other uses (like the study of mill
media sizing effects, described later), it has also been used
to model the Bond Work Index test (Ciribeni et al., 2020).
While it relies on the above-noted assumption, it avoids
the bewildering complexity, numerous assumptions, inac-
curacies associated with model fitting, and demonstrated
falsehoods of classical population balance-based computer
modelling and simulation (McIvor et al., 2023).
The same survey data as described above demonstrates
an integral link between the Functional Performance mill
grinding rate (MGrR) and the apparent cumulative mill
grinding rate (ACGR) at the same size of interest. They are,
in fact, different methods of calculating the same grinding
rate parameter, that is, the grinding rate of material desig-
nated as target plus size material per unit of energy applied
to target plus size material in the mill. The Functional
Performance mill grinding rate (a value calculated above
0.404 t/kWh) assumes that the percentage of plus size in the
mill is the average of that size in the mill feed and discharge.
The ACGR calculated at the same size (a value calculated
above 0.402 t/kWh) uses the same mill feed and discharge
size distributions and the same mill power but assumes
plug flow and first-order rate disappearance (i.e., constant
grinding rate per unit of plus size remaining as it dimin-
ishes in quantity). That they differ only to the third decimal
place for typical closed-circuit grinding mills attests to both
their fundamental difference and essential sameness. And
fortunately, used in conjunction with cyclone discrete size
separation performance and pump head and capacity per-
formance, this provides the ability to create a circuit mod-
elling program of high accuracy. Such knowledge of not
only how to model the mill reliably but also pump and
cyclone process application engineering provides a direct
step-by-step path to circuit optimization, as described in
the examples that follow.
EXAMPLES
Increasing Circuit Classification System Efficiency, CSE
The 2018 ball mill circuit survey at Fekola, referenced in
Clark (2023), showed a CSE of 76.1%. This compared to
values of CSE approximately 10 percentage points higher
in the Metcom database in circuits with fully optimized
pump and cyclone equipment. The circulating load ratio
was already suitably high (442%) and modern cyclones
with high sharpness of separation were being utilized. The
remaining opportunity was to improve the cyclone water
balance by increased feed water, subsequent reduction of
bypass of fines being returned to the mill, and therefore
increased CSE, as achieved in multiple cases in the past
(McIvor, 2014 McIvor et al, 2017).
Source :Clark, et al., 2023
Figure 3. Apparent cumulative grinding rate (by size class)—Fekola ball mill, 2018 circuit survey