3644 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
scientific basis for selection of this scale-up factor although
several practitioners have hypothesized the scale-up factor
to relate to power input and other dimensionless cell hydro-
dynamic parameters (Gorain et al, 2007 Grönstrand, et
al. 2014). With the recent trend towards larger flotation
cells with volumes of 600 m3 and greater in concentrators
along with the increasing adoption of non-mechanical cells
in flotation circuits, the problem of scale-up has become
even more acute. The scale-up factors that may have been
tolerated on smaller mechanical cells do not apply to larger
mechanical and non-mechanical cells. The transportation
of froth in these large cells adds further complexity to the
problem. Arbiter et al. (1980) foresaw this problem even
in the early days and pointed out that if scale-up were
attempted by rule-of-thumb and trial-and-error, then the
probability of failure would increase with scale.
There are several instances of underestimation of flota-
tion capacity due to improper selection of safety factors,
leading to operating plants not being capable of meeting
the metallurgical design target (Agar &Stratton-Crawley,
1982 Glatthaar et al, 2007 Klohn et al, 2016). This could
be a very risky proposition resulting in lost opportunities
due to production losses and the need for further capital
expenditure for additional flotation capacity. The reason for
high risks in conventional design is that the safety factors
are based on previous experiences on simpler ores, whereas
most of the present existing and future ore deposits are
metallurgically complex requiring a deeper understanding
of the mechanisms that drive flotation performance (Lane
et al., 2005).
This traditional approach has no doubt served the
industry very well in the past. But is still widely used by
various design engineers and engineering companies and it
is often difficult to discard due to its wide familiarity along
with the availability of historical data that allows some basic
benchmarking. Our experience has been that this approach
is still useful for a baseline design, which then must be vali-
dated and fine-tuned using other advanced techniques pre-
sented in this paper, leveraging the present know-how of
the various intricacies of the flotation process. This is the
main subject of discussion in this paper.
Modelling and Simulation Technique for Scale-Up and
Plant Design
Flotation modelling and simulation techniques have
emerged as an important toolset to provide a more rational
basis to design and optimize flotation circuits (Manlapig
et al., 1997 Gorain and Stradling, 2002 Harris et al.,
2002 Herbst and Harris, 2007 Dobby and Savassi, 2005
Harbort &Quan, 2017). Initiatives in this direction have
been driven by some of the major mining companies mainly
to increase the confidence level in designing flotation cir-
cuits with minimal risks. This also allows simulation of
“what-if” scenarios mainly to understand trade-off between
incremental recovery and flotation capacity requirements
or capital expenditure, and ultimately in optimization of
a flotation circuit design. Once a robust model has been
developed for a deposit, simulations can be done to under-
stand the effects of ore variability and circuit configura-
tions on flotation performance. Flotation modelling and
simulation techniques, although having a scientific basis,
are not perfect due to difficulties in modelling flotation
of complex ore types. The assumptions made for simula-
tions should be carefully examined for better confidence in
model predictions.
There are a few commercially available flotation simu-
lators, one is JKSimFloat developed by JKTech in Australia
and another is FLEET (Flotation Economic Evaluation
Tool) developed by Minnovex (now SGS) in Canada.
Both these techniques focus on the derivation of flotation
parameters that quantify the various flotation mechanisms
through bench scale test work on samples obtained from
plant, pilot plant or freshly prepared in the laboratory. The
flotation testing procedures and the derived parameters,
however, are different for these two techniques.
The JKSimFloat simulator has been considered for flo-
tation circuit design by the author because of its robustness
along with rigorous validation at various operations treat-
ing complex ore types. The JKSimFloat database is elabo-
rate with significant plant measurement data for rougher
and cleaner circuits along with extensive cell hydrodynamic
measurement data for various cell sizes and types including
non-mechanical cells (Coleman et al, 2006 Collins et al,
2009 Tabosa et al, 2020). This allows for enhanced param-
eter estimation with a higher level of confidence in predict-
ing flotation outcomes.
The JKSimFloat modeling is based on the model devel-
oped by the AMIRA P9 project (Schwarz and Richardson,
2013 Stange et al, 2014). The model separates the effects
of ore floatability, machine and froth behaviour from the
overall flotation rate constant (a lumped parameter), which
is explained below:
• Ore floatability (P): P is dependent on inherent float-
ability of the ore representing the mineral liberation
and flotation chemistry behaviour.
• Machine characteristics (Sb): Sb is a function of bub-
ble size generated by the flotation cell and superfi-
cial gas velocity (dependent on air flow rate and cell
cross-sectional area)
scientific basis for selection of this scale-up factor although
several practitioners have hypothesized the scale-up factor
to relate to power input and other dimensionless cell hydro-
dynamic parameters (Gorain et al, 2007 Grönstrand, et
al. 2014). With the recent trend towards larger flotation
cells with volumes of 600 m3 and greater in concentrators
along with the increasing adoption of non-mechanical cells
in flotation circuits, the problem of scale-up has become
even more acute. The scale-up factors that may have been
tolerated on smaller mechanical cells do not apply to larger
mechanical and non-mechanical cells. The transportation
of froth in these large cells adds further complexity to the
problem. Arbiter et al. (1980) foresaw this problem even
in the early days and pointed out that if scale-up were
attempted by rule-of-thumb and trial-and-error, then the
probability of failure would increase with scale.
There are several instances of underestimation of flota-
tion capacity due to improper selection of safety factors,
leading to operating plants not being capable of meeting
the metallurgical design target (Agar &Stratton-Crawley,
1982 Glatthaar et al, 2007 Klohn et al, 2016). This could
be a very risky proposition resulting in lost opportunities
due to production losses and the need for further capital
expenditure for additional flotation capacity. The reason for
high risks in conventional design is that the safety factors
are based on previous experiences on simpler ores, whereas
most of the present existing and future ore deposits are
metallurgically complex requiring a deeper understanding
of the mechanisms that drive flotation performance (Lane
et al., 2005).
This traditional approach has no doubt served the
industry very well in the past. But is still widely used by
various design engineers and engineering companies and it
is often difficult to discard due to its wide familiarity along
with the availability of historical data that allows some basic
benchmarking. Our experience has been that this approach
is still useful for a baseline design, which then must be vali-
dated and fine-tuned using other advanced techniques pre-
sented in this paper, leveraging the present know-how of
the various intricacies of the flotation process. This is the
main subject of discussion in this paper.
Modelling and Simulation Technique for Scale-Up and
Plant Design
Flotation modelling and simulation techniques have
emerged as an important toolset to provide a more rational
basis to design and optimize flotation circuits (Manlapig
et al., 1997 Gorain and Stradling, 2002 Harris et al.,
2002 Herbst and Harris, 2007 Dobby and Savassi, 2005
Harbort &Quan, 2017). Initiatives in this direction have
been driven by some of the major mining companies mainly
to increase the confidence level in designing flotation cir-
cuits with minimal risks. This also allows simulation of
“what-if” scenarios mainly to understand trade-off between
incremental recovery and flotation capacity requirements
or capital expenditure, and ultimately in optimization of
a flotation circuit design. Once a robust model has been
developed for a deposit, simulations can be done to under-
stand the effects of ore variability and circuit configura-
tions on flotation performance. Flotation modelling and
simulation techniques, although having a scientific basis,
are not perfect due to difficulties in modelling flotation
of complex ore types. The assumptions made for simula-
tions should be carefully examined for better confidence in
model predictions.
There are a few commercially available flotation simu-
lators, one is JKSimFloat developed by JKTech in Australia
and another is FLEET (Flotation Economic Evaluation
Tool) developed by Minnovex (now SGS) in Canada.
Both these techniques focus on the derivation of flotation
parameters that quantify the various flotation mechanisms
through bench scale test work on samples obtained from
plant, pilot plant or freshly prepared in the laboratory. The
flotation testing procedures and the derived parameters,
however, are different for these two techniques.
The JKSimFloat simulator has been considered for flo-
tation circuit design by the author because of its robustness
along with rigorous validation at various operations treat-
ing complex ore types. The JKSimFloat database is elabo-
rate with significant plant measurement data for rougher
and cleaner circuits along with extensive cell hydrodynamic
measurement data for various cell sizes and types including
non-mechanical cells (Coleman et al, 2006 Collins et al,
2009 Tabosa et al, 2020). This allows for enhanced param-
eter estimation with a higher level of confidence in predict-
ing flotation outcomes.
The JKSimFloat modeling is based on the model devel-
oped by the AMIRA P9 project (Schwarz and Richardson,
2013 Stange et al, 2014). The model separates the effects
of ore floatability, machine and froth behaviour from the
overall flotation rate constant (a lumped parameter), which
is explained below:
• Ore floatability (P): P is dependent on inherent float-
ability of the ore representing the mineral liberation
and flotation chemistry behaviour.
• Machine characteristics (Sb): Sb is a function of bub-
ble size generated by the flotation cell and superfi-
cial gas velocity (dependent on air flow rate and cell
cross-sectional area)