XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3 1489
(BA), the latter being the largest producer, accounting for
87.6%, and produced by Companhia de Ferro Ligas da
Bahia—FERBASA (Silva et al, 2022). FERBASA operates
in the Vale do Jacurici mineral district and accounts for
more than 90% of the national production of FeCr alloys.
Its main mines are in the municipalities of Andorinhas
(Ipueira mine) and Campo Formoso (Pedrinhas mine), BA.
The Cr ore is processed using gravity separation (spiral con-
centrators) in both mines and using sensor-based sensors
and jigs in the Ipueira mine. For that the ROM is crushed,
screened, and milled in a rod mill followed by a ball mill.
All production from the mine (concentrate and lump ore)
is destined for FERBASA metallurgical plant, located in
Pojuca, BA, where the FeCr alloy (high and low carbon) is
produced, as well as silicon-chrome, which are used in the
manufacture of special steels and stainless steel (Sampaio et
al. 2001).
According to Leroy et al. (2011), the d75 for chromite
flotation is commonly around 75 µm (200#). The round-
ness of the particles plays a significant role in flotation.
According to Guven et al. (2020), a reduction in the round-
ness from 0.797 to 0.732 results in an increase in chromite
flotation recovery, from 51% to 75.4% in microflotation
experiments. Bench scale flotation tests were performed to
decrease the silica content in FERBASAs chromite concen-
trate. However, it was possible to notice that the particle
size distribution and shape were not optimal for this end.
Therefore, the present work aims in analyze the particle size
distribution and morphological parameters (roundness, cir-
cularity, solidity, and aspect ratio) of a sample of chromite
concentrate to select the optimal milling conditions for this
material.
MATERIALS AND METHODS
Samples of chromite concentrate from FERBASA were
shipped to the Modelling and Mineral Processing Research
Lab (LaMPPMin) where they were dry oven dried at 100
°C for 24 hours. After that, the samples were homogenized
and quartered using Jones riffle splitter resulting in sub-
samples with 5, 7.5 and 10 kg. The milling tests were car-
ried out dry in a ball mill (30.5 cm diameter and 30 cm
height) filled with 14.9 kg of 27.5 mm steel balls. Two mill-
ing durations were tested (15 and 30 minutes). The mill
speed was kept constant at 72 rpm in all tests.
After milling the samples were quartered again and ali-
quots of approximately 500 g were wet sieved for 15 min-
utes, using at least 20 L of tap water per sieving, or until
the water passing through the last sieve presented no visible
particles or coloration (Sampaio 2007). A solution of 0,5%
(w/w) of sodium metasilicate (10 mL) was used as disper-
sant. After sieving the material retained in the sieves were
oven dried at 100 °C for 24 hours and the masses were
weighted.
The images acquisition to determine the morphological
parameters of the particles was performed using a Bresser
Biolux NV microscope with a HD digital camera (CCD)
coupled to it and CamLabLite software around 50 images
of the material retained in each sieve were acquired and
processed. The Image Processing &Image Analysis (IPIA)
were carried out using the ImageJ software. First a color
space conversion was performed to change from RGB (Red,
Green, and Blue) color space to YUV color space. YUV
means ‘luma,’ ‘red projection’ and ‘blue projection,’ which
is a mathematical encoding system that incorporates both
brightness and color. The YUV color space was defined in
this work as presented below:
Y =0.299 *R +0.587 *G +0.114 *B
U =0.492 *(B − Y)
V =0.877 *(R − Y)
After that the images were segmented, scaled, and the mor-
phological parameters were extracted. The particles ana-
lyzed ranged from 0.01 and 0.1 mm2, excluding particles
that touched the edges of the images, thus avoiding the
sub dimensioning of the particles. The segmentation of the
images was carried out transforming them to 8 bits images
and then using a threshold for binarization. The images
scale was adjusted to 811.25 pixels/mm.
The morphological parameters extracted of each par-
ticle were the area (measured of the internal space of the
particle), perimeter (length of the particle contour), circu-
larity (C), aspect ratio (AR), roundness (R), and solidity
(S), defined as:
C =4 *π *Area /(Perimeter)2
AR =Major Axis /Minor Axis
R =4 *Area /[π *(Major Axis)2 ]
S =Area /Convex Area
A Python script was written to batch process the digital
imagens. In total 2.053 images were processed. The result
of the process was saved as a csv file, one per digital image
containing the measurements of all particles found in the
image. Another Python script was written to merge all csv
(BA), the latter being the largest producer, accounting for
87.6%, and produced by Companhia de Ferro Ligas da
Bahia—FERBASA (Silva et al, 2022). FERBASA operates
in the Vale do Jacurici mineral district and accounts for
more than 90% of the national production of FeCr alloys.
Its main mines are in the municipalities of Andorinhas
(Ipueira mine) and Campo Formoso (Pedrinhas mine), BA.
The Cr ore is processed using gravity separation (spiral con-
centrators) in both mines and using sensor-based sensors
and jigs in the Ipueira mine. For that the ROM is crushed,
screened, and milled in a rod mill followed by a ball mill.
All production from the mine (concentrate and lump ore)
is destined for FERBASA metallurgical plant, located in
Pojuca, BA, where the FeCr alloy (high and low carbon) is
produced, as well as silicon-chrome, which are used in the
manufacture of special steels and stainless steel (Sampaio et
al. 2001).
According to Leroy et al. (2011), the d75 for chromite
flotation is commonly around 75 µm (200#). The round-
ness of the particles plays a significant role in flotation.
According to Guven et al. (2020), a reduction in the round-
ness from 0.797 to 0.732 results in an increase in chromite
flotation recovery, from 51% to 75.4% in microflotation
experiments. Bench scale flotation tests were performed to
decrease the silica content in FERBASAs chromite concen-
trate. However, it was possible to notice that the particle
size distribution and shape were not optimal for this end.
Therefore, the present work aims in analyze the particle size
distribution and morphological parameters (roundness, cir-
cularity, solidity, and aspect ratio) of a sample of chromite
concentrate to select the optimal milling conditions for this
material.
MATERIALS AND METHODS
Samples of chromite concentrate from FERBASA were
shipped to the Modelling and Mineral Processing Research
Lab (LaMPPMin) where they were dry oven dried at 100
°C for 24 hours. After that, the samples were homogenized
and quartered using Jones riffle splitter resulting in sub-
samples with 5, 7.5 and 10 kg. The milling tests were car-
ried out dry in a ball mill (30.5 cm diameter and 30 cm
height) filled with 14.9 kg of 27.5 mm steel balls. Two mill-
ing durations were tested (15 and 30 minutes). The mill
speed was kept constant at 72 rpm in all tests.
After milling the samples were quartered again and ali-
quots of approximately 500 g were wet sieved for 15 min-
utes, using at least 20 L of tap water per sieving, or until
the water passing through the last sieve presented no visible
particles or coloration (Sampaio 2007). A solution of 0,5%
(w/w) of sodium metasilicate (10 mL) was used as disper-
sant. After sieving the material retained in the sieves were
oven dried at 100 °C for 24 hours and the masses were
weighted.
The images acquisition to determine the morphological
parameters of the particles was performed using a Bresser
Biolux NV microscope with a HD digital camera (CCD)
coupled to it and CamLabLite software around 50 images
of the material retained in each sieve were acquired and
processed. The Image Processing &Image Analysis (IPIA)
were carried out using the ImageJ software. First a color
space conversion was performed to change from RGB (Red,
Green, and Blue) color space to YUV color space. YUV
means ‘luma,’ ‘red projection’ and ‘blue projection,’ which
is a mathematical encoding system that incorporates both
brightness and color. The YUV color space was defined in
this work as presented below:
Y =0.299 *R +0.587 *G +0.114 *B
U =0.492 *(B − Y)
V =0.877 *(R − Y)
After that the images were segmented, scaled, and the mor-
phological parameters were extracted. The particles ana-
lyzed ranged from 0.01 and 0.1 mm2, excluding particles
that touched the edges of the images, thus avoiding the
sub dimensioning of the particles. The segmentation of the
images was carried out transforming them to 8 bits images
and then using a threshold for binarization. The images
scale was adjusted to 811.25 pixels/mm.
The morphological parameters extracted of each par-
ticle were the area (measured of the internal space of the
particle), perimeter (length of the particle contour), circu-
larity (C), aspect ratio (AR), roundness (R), and solidity
(S), defined as:
C =4 *π *Area /(Perimeter)2
AR =Major Axis /Minor Axis
R =4 *Area /[π *(Major Axis)2 ]
S =Area /Convex Area
A Python script was written to batch process the digital
imagens. In total 2.053 images were processed. The result
of the process was saved as a csv file, one per digital image
containing the measurements of all particles found in the
image. Another Python script was written to merge all csv