2854 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
• Recirculation load: without and with 6 m3/h (ca.
30% recirculation load).
• Reagents: a mixture of ethyl and isobutyl sodium
xanthates as collector and polyglycol alkyl ethers
used as frother.
AUTOMATED MINERALOGICAL
ANALYSIS (MLA)
The composite feed, concentrate and tailings were collected
and split into a few grams with rotary splitters. All the sam-
ples were carbon coated using a Leica MED 020 vacuum
evaporator to ensure the conductivity of the sample sur-
face. Note that the ore samples also contain organic mat-
ter which was unfortunately not taken into account when
preparing the sample for MLA analysis. Grain mounts are
prepared as B-sections to avoid issues with sedimentation,
following the routine proposed by Heinig et al. (2015).
After carbon coating, MLA measurements are conducted at
HIF in a FEI Quanta 650F scanning electron microscope
equipped with two Bruker Quantax X-Flash 5030 energy
dispersive X-ray spectrometry detectors and MLA Suite
3.1.4 for automated data acquisition. Grain-based X-ray
mapping measurements are conducted with 1 µm pixel
size, 6 µm X-ray spacing, 25 kV operating voltage, and a
minimum particle area of 2 pixels.
PARTICLE-BASED MODELS
A particle-based model (Pereira et al., 2021) is trained for
each flotation test with its processing products. The PSM is
a least absolute shrinkage and selection operator (LASSO)-
regularized logistic regression, fed with individual particle
data obtained with automated mineralogy: size (equivalent
circle diameter, ECD), shape (aspect ratio and solidity),
modal composition, and surface composition. Data trans-
formation consists of:
• Log transformation of size and shape variables,
• The square of the log-transformed ECD is added as
a new variable,
• Modal and surface composition vectors are individu-
ally closed to sum to 100% (van den Boogaart and
Tolosana-Delgado, 2013),
• Adding a categorical variable indicating the main
mineral composing the particle (in mass) to the data-
set—namely mainmineral.
Particle data, compiled in a training dataset that contains
the same number of particles for each process stream, is
used to train the PSM. From all particle data, 70% is used
for model training and the remaining 30% for validation.
The response variable of the logistic regression is particle
class (i.e., concentrate or tailings). The model allows for
Figure 3. Mineral composition of feed sample of four different pilot tests
• Recirculation load: without and with 6 m3/h (ca.
30% recirculation load).
• Reagents: a mixture of ethyl and isobutyl sodium
xanthates as collector and polyglycol alkyl ethers
used as frother.
AUTOMATED MINERALOGICAL
ANALYSIS (MLA)
The composite feed, concentrate and tailings were collected
and split into a few grams with rotary splitters. All the sam-
ples were carbon coated using a Leica MED 020 vacuum
evaporator to ensure the conductivity of the sample sur-
face. Note that the ore samples also contain organic mat-
ter which was unfortunately not taken into account when
preparing the sample for MLA analysis. Grain mounts are
prepared as B-sections to avoid issues with sedimentation,
following the routine proposed by Heinig et al. (2015).
After carbon coating, MLA measurements are conducted at
HIF in a FEI Quanta 650F scanning electron microscope
equipped with two Bruker Quantax X-Flash 5030 energy
dispersive X-ray spectrometry detectors and MLA Suite
3.1.4 for automated data acquisition. Grain-based X-ray
mapping measurements are conducted with 1 µm pixel
size, 6 µm X-ray spacing, 25 kV operating voltage, and a
minimum particle area of 2 pixels.
PARTICLE-BASED MODELS
A particle-based model (Pereira et al., 2021) is trained for
each flotation test with its processing products. The PSM is
a least absolute shrinkage and selection operator (LASSO)-
regularized logistic regression, fed with individual particle
data obtained with automated mineralogy: size (equivalent
circle diameter, ECD), shape (aspect ratio and solidity),
modal composition, and surface composition. Data trans-
formation consists of:
• Log transformation of size and shape variables,
• The square of the log-transformed ECD is added as
a new variable,
• Modal and surface composition vectors are individu-
ally closed to sum to 100% (van den Boogaart and
Tolosana-Delgado, 2013),
• Adding a categorical variable indicating the main
mineral composing the particle (in mass) to the data-
set—namely mainmineral.
Particle data, compiled in a training dataset that contains
the same number of particles for each process stream, is
used to train the PSM. From all particle data, 70% is used
for model training and the remaining 30% for validation.
The response variable of the logistic regression is particle
class (i.e., concentrate or tailings). The model allows for
Figure 3. Mineral composition of feed sample of four different pilot tests