2872 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
added through a distribution plate located at top of the
reverse fluidized bed. The concentrate left the cell as an
overflow while the underflow was pumped to the tailing
tank through a discharge port at bottom of the cell.
For each test, 250 g of chalcopyrite sample was freshly
ground to P80 of 106 μm. It was then blended with 4.75 kg
of quartz powder to produce a head grade of 5% chalcopy-
rite. The system was first run for 5 min to ensure it reached
an equilibrium. Then, four sets of product samples were
taken at 5, 10, 15 and 20 min. Each set contained two
samples from the concentrate and tailing pipes. To mini-
mize the influence from system fluctuation, each sample
was taken for 3 min. The samples were then dried and pul-
verized for ICP-MS analysis to determine copper concen-
tration for calculating mineral recoveries.
The operating parameters for the baseline test, includ-
ing the feed solid concentration, gas flux, feed flux, wash
water flux and the tailing discharged rate were based on
those optimized for coal flotation (Chen et al., 2023).
Then, the gas flux and the wash water flux were adjusted to
minimize gangue entrainment in chalcopyrite flotation. All
the conditions tested are summarized in Table 1, where the
gas flux and feed flux were calculated based on the cross-
sectional area of the downcomer while the wash water flux
and tailing discharge rate were calculated based on the ver-
tical cell cross-sectional area.
CFD Simulation
CFD simulation was performed with the open-source
packages, OpenFOAM. As the gas fraction and the liquid
motion are the main interest in this study, a two-phase
gas-liquid CFD simulation was performed. To simulate
chalcopyrite particle-laden bubbles, the relative density of
gas phase was defined based on the relative density of chal-
copyrite particle-bubble aggregates. A typical chalcopyrite
loading on the bubbles in industrial flotation cells, around
26.82 g/L gas (Moys et al., 2010), was used to estimate the
relative density of chalcopyrite particle-bubble aggregates.
As the main interest of this study was gangue entrain-
ment, the CFD simulation was focused on the reverse fluid-
ized bed, which is the major region responsible for gangue
reduction in the RFC. Hence, the geometry model used in
the simulation adapted the actual parameter of the reverse
fluidized bed, while the downcomer was simplified and not
simulated in this study. The inclined channel was also sim-
plified into a large vertical channel to prevent the loss of
bubbles to the tailing during the simulation.
The solver used in the CFD simulation was built based
on the in-built solver, multiphaseEulerFoam. The turbu-
lence model (mixture k-epsilon (k-ε) model) was chosen
for the solution of energy balance between liquid and
gas phases to better simulate the formation of turbulence
within the RFC. Force models for major interphase forces,
M
Concentrate
Wash
Water
Compressed
Air
Tailing
Feed
Figure 2. Layout for RFC set-up (Chen et al., 2023)
added through a distribution plate located at top of the
reverse fluidized bed. The concentrate left the cell as an
overflow while the underflow was pumped to the tailing
tank through a discharge port at bottom of the cell.
For each test, 250 g of chalcopyrite sample was freshly
ground to P80 of 106 μm. It was then blended with 4.75 kg
of quartz powder to produce a head grade of 5% chalcopy-
rite. The system was first run for 5 min to ensure it reached
an equilibrium. Then, four sets of product samples were
taken at 5, 10, 15 and 20 min. Each set contained two
samples from the concentrate and tailing pipes. To mini-
mize the influence from system fluctuation, each sample
was taken for 3 min. The samples were then dried and pul-
verized for ICP-MS analysis to determine copper concen-
tration for calculating mineral recoveries.
The operating parameters for the baseline test, includ-
ing the feed solid concentration, gas flux, feed flux, wash
water flux and the tailing discharged rate were based on
those optimized for coal flotation (Chen et al., 2023).
Then, the gas flux and the wash water flux were adjusted to
minimize gangue entrainment in chalcopyrite flotation. All
the conditions tested are summarized in Table 1, where the
gas flux and feed flux were calculated based on the cross-
sectional area of the downcomer while the wash water flux
and tailing discharge rate were calculated based on the ver-
tical cell cross-sectional area.
CFD Simulation
CFD simulation was performed with the open-source
packages, OpenFOAM. As the gas fraction and the liquid
motion are the main interest in this study, a two-phase
gas-liquid CFD simulation was performed. To simulate
chalcopyrite particle-laden bubbles, the relative density of
gas phase was defined based on the relative density of chal-
copyrite particle-bubble aggregates. A typical chalcopyrite
loading on the bubbles in industrial flotation cells, around
26.82 g/L gas (Moys et al., 2010), was used to estimate the
relative density of chalcopyrite particle-bubble aggregates.
As the main interest of this study was gangue entrain-
ment, the CFD simulation was focused on the reverse fluid-
ized bed, which is the major region responsible for gangue
reduction in the RFC. Hence, the geometry model used in
the simulation adapted the actual parameter of the reverse
fluidized bed, while the downcomer was simplified and not
simulated in this study. The inclined channel was also sim-
plified into a large vertical channel to prevent the loss of
bubbles to the tailing during the simulation.
The solver used in the CFD simulation was built based
on the in-built solver, multiphaseEulerFoam. The turbu-
lence model (mixture k-epsilon (k-ε) model) was chosen
for the solution of energy balance between liquid and
gas phases to better simulate the formation of turbulence
within the RFC. Force models for major interphase forces,
M
Concentrate
Wash
Water
Compressed
Air
Tailing
Feed
Figure 2. Layout for RFC set-up (Chen et al., 2023)