2032 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
et al., 2016). The particle velocity at the inlet was set to
be identical to the fluid, with interaction between the par-
ticles assumed to be negligible. In this Lagrangian model,
the velocity distribution of particles can be evaluated by the
force balances on the particle, as given in (6):
F u
g^t
p
D p
pt
pt
2t
2u
t
th
=-
-
^u h+ (6)
In which:
up =Particle velocity
rpt =Particle density
g =Gravitational acceleration
FD =Drag force per unit particle mass
For the drag force calculation, the standard spherical
drag law was used. A summary of the properties utilized
for Eulerian phases and the six distinct particle injections
for the discrete particle phases is given in Table 2. Rosin
Rammler distributions were used for the ranges given.
RESULTS AND DISCUSSION
HLS Test Results
Densimetric analysis was done at six relative densities of
2.3, 2.5, 2.6, 2.65 and 2.96 g/cm3. Figure 4 represents
the data obtained from HLS washability tests. These high-
lighted that that the total sample mass was contained
mainly at three densities of 2.65, 2.6 and 2.3 g/cm3 with
negligible masses at others. For P2O5, the sample showed
a bi-modal distribution at 2.65 and 2.3 g/cm3, meaning
that near density material at cut points between these values
would present challenges in separation. The lower densities
than expected for apatite/flourapatite (theoretically 3.1–3.2
g/cm3) was likely due to high porosity within the grains.
SiO2 was predominantly present at a density of 2.6 g/cm3,
at which it accounted for almost all material, with a sig-
nificant amount present also at 2.65 g/cm3 and a much
lower quantity at 2.3 g/cm3. P2O5 upgrade was therefore
favourable at density cut points of 2.3 and 2.65 g/cm3. At
2.3 g/cm3, a grade of 29.9% P2O5 was achieved, which was
slightly below the target specification of 31%.
Modelled HLS Data for Pilot DMS Performance
Prediction
The modelled DMS mass yields and product grades at dif-
ferent cut points are given in Figure 5. After incorporating
inefficiencies for the yield, it was identified that a density
cut point of 2.5g/cm3 would achieve similar mass yields
and P2O5 grades to those obtained during laboratory test-
ing. At a cut point of 2.5g/cm3, the anticipated mass yield
Figure 3. Dense medium cyclone (Upper left), CAD model (Upper right) and meshed geometry utilized for CFD simulations
Table 1. Summary of CFD settings utilized with RSM
Calculation Solution method
Pressure-Velocity scheme PISO
Pressure PRESTO!
Momentum Second order upwind
Turbulent kinetic energy Second order upwind
Turbulent dissipation rate Second order upwind
Reynold’s Stresses Second order upwind
Transient formulation First order implicit
et al., 2016). The particle velocity at the inlet was set to
be identical to the fluid, with interaction between the par-
ticles assumed to be negligible. In this Lagrangian model,
the velocity distribution of particles can be evaluated by the
force balances on the particle, as given in (6):
F u
g^t
p
D p
pt
pt
2t
2u
t
th
=-
-
^u h+ (6)
In which:
up =Particle velocity
rpt =Particle density
g =Gravitational acceleration
FD =Drag force per unit particle mass
For the drag force calculation, the standard spherical
drag law was used. A summary of the properties utilized
for Eulerian phases and the six distinct particle injections
for the discrete particle phases is given in Table 2. Rosin
Rammler distributions were used for the ranges given.
RESULTS AND DISCUSSION
HLS Test Results
Densimetric analysis was done at six relative densities of
2.3, 2.5, 2.6, 2.65 and 2.96 g/cm3. Figure 4 represents
the data obtained from HLS washability tests. These high-
lighted that that the total sample mass was contained
mainly at three densities of 2.65, 2.6 and 2.3 g/cm3 with
negligible masses at others. For P2O5, the sample showed
a bi-modal distribution at 2.65 and 2.3 g/cm3, meaning
that near density material at cut points between these values
would present challenges in separation. The lower densities
than expected for apatite/flourapatite (theoretically 3.1–3.2
g/cm3) was likely due to high porosity within the grains.
SiO2 was predominantly present at a density of 2.6 g/cm3,
at which it accounted for almost all material, with a sig-
nificant amount present also at 2.65 g/cm3 and a much
lower quantity at 2.3 g/cm3. P2O5 upgrade was therefore
favourable at density cut points of 2.3 and 2.65 g/cm3. At
2.3 g/cm3, a grade of 29.9% P2O5 was achieved, which was
slightly below the target specification of 31%.
Modelled HLS Data for Pilot DMS Performance
Prediction
The modelled DMS mass yields and product grades at dif-
ferent cut points are given in Figure 5. After incorporating
inefficiencies for the yield, it was identified that a density
cut point of 2.5g/cm3 would achieve similar mass yields
and P2O5 grades to those obtained during laboratory test-
ing. At a cut point of 2.5g/cm3, the anticipated mass yield
Figure 3. Dense medium cyclone (Upper left), CAD model (Upper right) and meshed geometry utilized for CFD simulations
Table 1. Summary of CFD settings utilized with RSM
Calculation Solution method
Pressure-Velocity scheme PISO
Pressure PRESTO!
Momentum Second order upwind
Turbulent kinetic energy Second order upwind
Turbulent dissipation rate Second order upwind
Reynold’s Stresses Second order upwind
Transient formulation First order implicit