XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3 919
used as SW). One important limit is the required dilution
of input water by SW for loss compensation, which repre-
sent around 30% of input water. As the experiment pro-
gresses, the concentration increases. As the same amount
of water is lost after each cycle, more and more sulfates are
lost. At some point, the amount lost equals the amount
dissolved, and equilibrium is reached. This is closer to a
pseudo-equilibrium related to the test residence time and
water compensation dilution ratio than a saturation due to
thermodynamical equilibrium.
Even if it is not too bad, the simultaneous fit on DW
and PW is not as suitable for magnesium and calcium, even
at very high Cmax. Fit was thus made only on PW water
(values are summed up in Table 3), which is considered
more representative for planned simulations as it might
be closer to the expected recirculated water that would be
obtained in closed loop. For calcium, it has been considered
that gypsum was the main form of sulfate precipitation:
stoichiometry was imposed on Cmax and only k was fitted,
The integration of equation 1 over the residence time
distribution of a perfect mixer leads to the steady-state
model used for plant simulation, mill and flotation cells
being modelled as a Continuous Stirred Tank Reactor
(CSTR):
C C C C kx
kx
1 max output input input =+-+^h (4)
General Plant
Recirculation Loop Without Water Treatment
The global plant model on which this work is based is
described in Braak et al. (2022), including details on flota-
tion and water treatment operations modelling. This model
was slightly modified to assess a fictive recirculation loop
scenario. Only the Cu circuit is considered here with an
added Solid/Liquid (S/L) separation between Cu and Ni
circuits and water recirculation into the Cu circuit. The
simplified flowsheet (cleaner section is not detailed) is pre-
sented in Figure 5.
Dissolution is simulated by associating a chemical reac-
tor model to the grinding stage, as well as to rougher and
rougher/scavenger flotation stage models. It is not consid-
ered for the cleaner section, as dissolution tests have been
made for rougher and rougher/scavenger stage conditions.
pH is higher in the cleaner section which may impact sol-
ubility and precipitation behavior. It is assumed that the
plant operating conditions (pH, depressant…) prevails in
this section: flotation kinetics is kept constant, and the
influence of water composition variation is neglected. For
similar reason, another simplification in the flowsheet is
made regarding the recirculation of process water from con-
centrate thickening which is not considered here. Another
reason for assuming to neglect the impact of dissolution
in cleaner section is the fact that the greatest fraction of
Figure 4. Comparison between process water (PW) and distilled water (DW) tests
for sulfates at 20°C
Table 3. Calibrated parameters for dissolution test (magnesium and calcium)
Fit C
max ,mg/L k, /min Error
PW Mg 340 1.65 10–3 27.7
PW Ca 1 110 4.42 10–4 19.4
used as SW). One important limit is the required dilution
of input water by SW for loss compensation, which repre-
sent around 30% of input water. As the experiment pro-
gresses, the concentration increases. As the same amount
of water is lost after each cycle, more and more sulfates are
lost. At some point, the amount lost equals the amount
dissolved, and equilibrium is reached. This is closer to a
pseudo-equilibrium related to the test residence time and
water compensation dilution ratio than a saturation due to
thermodynamical equilibrium.
Even if it is not too bad, the simultaneous fit on DW
and PW is not as suitable for magnesium and calcium, even
at very high Cmax. Fit was thus made only on PW water
(values are summed up in Table 3), which is considered
more representative for planned simulations as it might
be closer to the expected recirculated water that would be
obtained in closed loop. For calcium, it has been considered
that gypsum was the main form of sulfate precipitation:
stoichiometry was imposed on Cmax and only k was fitted,
The integration of equation 1 over the residence time
distribution of a perfect mixer leads to the steady-state
model used for plant simulation, mill and flotation cells
being modelled as a Continuous Stirred Tank Reactor
(CSTR):
C C C C kx
kx
1 max output input input =+-+^h (4)
General Plant
Recirculation Loop Without Water Treatment
The global plant model on which this work is based is
described in Braak et al. (2022), including details on flota-
tion and water treatment operations modelling. This model
was slightly modified to assess a fictive recirculation loop
scenario. Only the Cu circuit is considered here with an
added Solid/Liquid (S/L) separation between Cu and Ni
circuits and water recirculation into the Cu circuit. The
simplified flowsheet (cleaner section is not detailed) is pre-
sented in Figure 5.
Dissolution is simulated by associating a chemical reac-
tor model to the grinding stage, as well as to rougher and
rougher/scavenger flotation stage models. It is not consid-
ered for the cleaner section, as dissolution tests have been
made for rougher and rougher/scavenger stage conditions.
pH is higher in the cleaner section which may impact sol-
ubility and precipitation behavior. It is assumed that the
plant operating conditions (pH, depressant…) prevails in
this section: flotation kinetics is kept constant, and the
influence of water composition variation is neglected. For
similar reason, another simplification in the flowsheet is
made regarding the recirculation of process water from con-
centrate thickening which is not considered here. Another
reason for assuming to neglect the impact of dissolution
in cleaner section is the fact that the greatest fraction of
Figure 4. Comparison between process water (PW) and distilled water (DW) tests
for sulfates at 20°C
Table 3. Calibrated parameters for dissolution test (magnesium and calcium)
Fit C
max ,mg/L k, /min Error
PW Mg 340 1.65 10–3 27.7
PW Ca 1 110 4.42 10–4 19.4