1568 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
blocks. This included 10,478 blocks of compact itabirite
and 10,509 blocks of friable itabirite. For ore types not
addressed by the statistical models, median values from
process tests were assigned. The mean Mass Recovery values
assigned to the blocks averaged at 37.3%, with a median of
38.3%. Mass recovery values ranged from a minimum of
20.7% to a maximum of 50.3%, with blocks showing lower
recovery predominantly concentrated in the southeastern
region of the model. Regarding Metallurgical Recovery, the
average values allocated to the blocks reached 62.6%, with
a median of 62.9%. The spectrum ranged from a minimum
of 49.2% to a maximum of 72.0%, with blocks exhibit-
ing lower metallurgical recovery values also clustered in the
southeastern part of the model (Figure 10).
In terms of energy requirement, the mean value found
for the blocks is 6.6 kWh/t, with a median of 7.0 kWh/t.
The minimum recorded value was 0.3 kWh/t, while the
maximum reached 9.2 kWh/t. Blocks with higher energy
requirement values are predominantly situated in the
northeast region of the model. Regarding SiO2 content in
the concentrate, the average value observed for the blocks
is 2.2%, with a median of 2.0%. The lowest recorded value
was 1.4%, while the highest was 16.4%. Blocks with higher
silica content in the concentrate are primarily located in the
southeast region of the model, as seen in Figure 11.
Validation of the Geometallurgical Model and
Comparison with Industrial Results
For the comparative analysis of predicted mass and metal-
lurgical recoveries in the model against industrial outcomes,
moments involving direct feed, specifically bypassing the
homogenization stockpile, were excluded. During these
intervals, no discernible patterns or trends were evident in
the data for mass and metallurgical recovery, likely due to
inadequate material homogenization, thereby complicat-
ing operational oversight. Moreover, stockpiles were omit-
ted if the sampled Fe grade at the homogenization yard
deviated by more than 6.5 percentage points from the Fe
grade projected by the geological model. In such instances,
it was assumed that the physical-chemical model of the
fronts was not congruent, resulting in the geometallurgi-
cal model data also being incongruent. The results of the
comparative analysis between predicted mass and metal-
lurgical recovery values in the geometallurgical model and
those observed industrially are represented in Figure 12 and
13, corresponding to the initial assessment period (August
Figure 10. Horizontes spatial geometallurgical models for mass and metallurgical
recoveries, respectively
blocks. This included 10,478 blocks of compact itabirite
and 10,509 blocks of friable itabirite. For ore types not
addressed by the statistical models, median values from
process tests were assigned. The mean Mass Recovery values
assigned to the blocks averaged at 37.3%, with a median of
38.3%. Mass recovery values ranged from a minimum of
20.7% to a maximum of 50.3%, with blocks showing lower
recovery predominantly concentrated in the southeastern
region of the model. Regarding Metallurgical Recovery, the
average values allocated to the blocks reached 62.6%, with
a median of 62.9%. The spectrum ranged from a minimum
of 49.2% to a maximum of 72.0%, with blocks exhibit-
ing lower metallurgical recovery values also clustered in the
southeastern part of the model (Figure 10).
In terms of energy requirement, the mean value found
for the blocks is 6.6 kWh/t, with a median of 7.0 kWh/t.
The minimum recorded value was 0.3 kWh/t, while the
maximum reached 9.2 kWh/t. Blocks with higher energy
requirement values are predominantly situated in the
northeast region of the model. Regarding SiO2 content in
the concentrate, the average value observed for the blocks
is 2.2%, with a median of 2.0%. The lowest recorded value
was 1.4%, while the highest was 16.4%. Blocks with higher
silica content in the concentrate are primarily located in the
southeast region of the model, as seen in Figure 11.
Validation of the Geometallurgical Model and
Comparison with Industrial Results
For the comparative analysis of predicted mass and metal-
lurgical recoveries in the model against industrial outcomes,
moments involving direct feed, specifically bypassing the
homogenization stockpile, were excluded. During these
intervals, no discernible patterns or trends were evident in
the data for mass and metallurgical recovery, likely due to
inadequate material homogenization, thereby complicat-
ing operational oversight. Moreover, stockpiles were omit-
ted if the sampled Fe grade at the homogenization yard
deviated by more than 6.5 percentage points from the Fe
grade projected by the geological model. In such instances,
it was assumed that the physical-chemical model of the
fronts was not congruent, resulting in the geometallurgi-
cal model data also being incongruent. The results of the
comparative analysis between predicted mass and metal-
lurgical recovery values in the geometallurgical model and
those observed industrially are represented in Figure 12 and
13, corresponding to the initial assessment period (August
Figure 10. Horizontes spatial geometallurgical models for mass and metallurgical
recoveries, respectively