2015 MW Supervisor Flotation
2015 MW Supervisor Flotation
2015 MW Supervisor Flotation
process simulator
Marie-Véronique DURANCE, Managing Director, Caspeo
Evgeny Chernykh, Managing Director, Computing System
Simulation is now widely used in many process industries. This paper presents the interest of process
simulation for studying the sensitivity of mineral processing plants. USIM PAC has been developed
since 1986 for the Mineral Processing sector. USIM PAC handles the design and optimization of entire
processing plants with a comprehensive set of mathematical models for unit operations that span from
crushing to refining. A variety of objective functions and tools, such as the simulation supervisor,
facilitate the global multi-criterion optimization. The supervisor of simulation uses the automation
concepts of sensors and actuators. At any place of the flowsheet, the user implements a “soft actuator”
by choosing the process parameter such as equipment size or settings, stream flowrate, percent solids
or d80, which can vary. Then, he can evaluate the effect of the variation of this value at any other place
of the flowsheet by inserting a “soft sensor”. Thus, by example, it becomes possible to look at the
evolution of the recoveries and/or the circulating loads when scanning an interval of the possible feed
flowrates.
Initial situation. The example presented here is built from several sets of plant data coming from
different plants and compiled to obtain realistic demonstration data. The flowsheet of the plant is
presented in Figure 1. It is a classical grinding flotation circuit processing a gold ore for producing a
gold concentrate. At the initial stage the flotation is fed by the hydrocyclones overflow without
regulation of the pulp density. The possibility to regulate the flotation feed density has been added as
an option.
Grinding flotation circuit of a gold ore
With percent solids
regulation
10 12
Ore feed Rougher
Water addition Final
flotation feed
1 3 7 6 tails
2 11
5
9 13
4
6 7 14
1 5
2 8
4 Without regulationConditioner
3
Rougher Rougher
15
concentrate
Hydrocyclone
Cleaner
Tails 18
8 16
9
Conditioner
17
Concentrate
19
10
Cleaner
The principle is to modify some of the operating parameters and the number of cells in the flotation
circuit and to check the impact on the grade/recovery curve.
Main equipment sizes are given in Table 1. The circuit feed size distribution is given in Figure 2.
Table 1. Main equipment sizes
In the initial configuration, with the operating parameters given in Table 2, the concentrate has a grade
of 121 g/t Au with a recovery of 80.4 %. Detailed performances are given in table 3.
d80 (µm)
100.00
75.00
90.00
70.00
80.00
65.00 70.00
60.00 60.00
25.00 30.00 35.00 40.00 45.00 50.00
% solids of hydrocyclone feed
The increase of pulp density for feeding hydrocyclones reduces the pressure drop thus increases the
d80 of the flotation feed (see Figure 5). However the overall gold recovery is not significantly impacted,
showing that this range of size is around the optimal liberation size.
Another supervisor cycle is run without regulating the flotation feed density. It appears that the
recovery is increasing when the feed pulp density increases (see Figure 6). This is due to a lower dilution
thus a higher residence time of the pulp in the flotation cells. The residence time is then the limiting
factor of the circuit.
90
80
90
70
d80 (µm)
60
50 80
40
30
70
20
10
0 60
25 30 35 40 45 50
% solids of hydrocyclone feed
Au recovery (%) % solids of cyclone OF (%) d80 of cyclone OF (µm)
85
20 rougher
3 cleaning cells
cells
Au recovery (%)
80 4 cleaning cells
5 cleaning cells
75
6 cleaning cells
70 7 cleaning cells
In Figure 7, each curve corresponds to a given number of cleaning cells (from 3 to 10). The different
points of a curve correspond to different numbers of rougher cells (from 10 to 20). It appears, as
anticipated, that the recovery largely depends on the number of roughing cells, while the evolution of
grades is more dependent on the number of cleaning cells. The economic impact has then to be
estimated considering the investment and operating costs of additional cells and the revenue
generated by higher recoveries and grades.
Conclusion. This paper shows the interest of simulation for exploring the sensitivity of a circuit. If the
final decision has to be taken considering economic factors, the technical elements provided by such
a tool largely assists the engineers in the selection of the best options.