5
• Box plots were reviewed between composites and the
model for the rest of the elements, where similarly,
the trend between them remains consistent.
It should be noted that it is possible to have FeM val-
ues in a RC but no values for the other elements. Since
FeM is used as the control grade and its kriging variance
as the control item, interpolated blocks will be obtained
with the additional 7 items, where there will be RCs around
those blocks that do not have values for these 7 items. These
blocks will be interpolated by the farthest RCs that have
FeM values. If this is not acceptable, separate runs should
also be performed, saving the kriging variance for each item
separately, and then resetting that item based on a deter-
mined variance value for each independent item (like how
values are reset with VAR 0.6).
5. Calcs in block model.
In this step are calculate the following variables:
• Using classification criteria, the results are saved in
the CLASS item.
• The specific gravity calculation is derived from a
regression formula and is saved in an SG item. With
the interpolated data of FEM and the specific gravity
calculation.
• The mineral concentrate is then calculated using the
mineral recovery formula provided by the mine’s pro-
cessing area, and the result is stored in the TCONC
item.
6. Interpolation of zones.
When there is a change in waste lithology in the area due to
mapping or updating of the lithological model, the specific
gravity of the lithology is assigned. This specific gravity is
determined through density sampling in the model.
Validation of ore control model
An exploratory data analysis (EDA) was conducted on
the reverse circulation sampling (EDA 2016) and the
data obtained from the Ore Control model (Implemented
2023). The comparison was made through a histogram and
a tonnage-grade curve. (Figure 8 and Figure 9)
The Table 5 shows an adjustment in the means is
observed due to the implementation of the model, showing
a slight smoothing effect that is not a significant impact.
Table 3. Composite statistics
Valid Min Max Mean
Standard
Desviation
Variance
Coefficient of
Variation
FEM 11,438 0.000 62.300 20.614 11.220 125.910 0.544
FET 8,641 56.110 71.000 67.682 1.700 2.921 0.025
Al2O3 3,629 0.000 3.400 0.910 0.430 0.189 0.475
CAO 8,641 0.210 6.250 0.760 0.314 0.098 0.411
MGO 3,629 0.000 2.700 0.375 0.249 0.062 0.661
P 8,641 0.000 0.114 0.015 0.111 0.000 0.720
S 8,641 0.000 2.760 0.120 0.167 0.028 1.394
SIO2 8,641 0.000 8.740 2.492 0.980 0.972 0.396
Table 4. Model statistics
Valid Min Max Mean Standard
Desviation
Variance Coefficient of
Variation
FEM 30,538 0.000 60.110 19.446 9.897 97.944 0.509
FET 30,306 57.540 70.610 67.449 1.564 2.445 0.023
Al2O3 16,402 0.000 3.072 0.881 0.378 0.143 0.429
CAO 30,306 0.363 4.699 0.795 0.296 0.088 0.372
MGO 16,402 0.000 2.103 0.352 0.205 0.042 0.581
P 30,306 0.000 0.114 0.017 0.011 0.001 0.642
S 30,306 0.000 2.287 0.131 0.152 0.023 1.162
SIO2 30,306 0.926 7.714 2.599 0.905 0.819 0.348
Figure 8. Histogram of the 2016 database (EDA
2016 in red) vs. the database after improvement in 2023
(Implemented 2023 in green)
Figure 9. Graph of tonnage-curve vs. grade (EDA 2016 in
green) compared to the database after improvement in 2023
(Implemented 2023 in blue)
• Box plots were reviewed between composites and the
model for the rest of the elements, where similarly,
the trend between them remains consistent.
It should be noted that it is possible to have FeM val-
ues in a RC but no values for the other elements. Since
FeM is used as the control grade and its kriging variance
as the control item, interpolated blocks will be obtained
with the additional 7 items, where there will be RCs around
those blocks that do not have values for these 7 items. These
blocks will be interpolated by the farthest RCs that have
FeM values. If this is not acceptable, separate runs should
also be performed, saving the kriging variance for each item
separately, and then resetting that item based on a deter-
mined variance value for each independent item (like how
values are reset with VAR 0.6).
5. Calcs in block model.
In this step are calculate the following variables:
• Using classification criteria, the results are saved in
the CLASS item.
• The specific gravity calculation is derived from a
regression formula and is saved in an SG item. With
the interpolated data of FEM and the specific gravity
calculation.
• The mineral concentrate is then calculated using the
mineral recovery formula provided by the mine’s pro-
cessing area, and the result is stored in the TCONC
item.
6. Interpolation of zones.
When there is a change in waste lithology in the area due to
mapping or updating of the lithological model, the specific
gravity of the lithology is assigned. This specific gravity is
determined through density sampling in the model.
Validation of ore control model
An exploratory data analysis (EDA) was conducted on
the reverse circulation sampling (EDA 2016) and the
data obtained from the Ore Control model (Implemented
2023). The comparison was made through a histogram and
a tonnage-grade curve. (Figure 8 and Figure 9)
The Table 5 shows an adjustment in the means is
observed due to the implementation of the model, showing
a slight smoothing effect that is not a significant impact.
Table 3. Composite statistics
Valid Min Max Mean
Standard
Desviation
Variance
Coefficient of
Variation
FEM 11,438 0.000 62.300 20.614 11.220 125.910 0.544
FET 8,641 56.110 71.000 67.682 1.700 2.921 0.025
Al2O3 3,629 0.000 3.400 0.910 0.430 0.189 0.475
CAO 8,641 0.210 6.250 0.760 0.314 0.098 0.411
MGO 3,629 0.000 2.700 0.375 0.249 0.062 0.661
P 8,641 0.000 0.114 0.015 0.111 0.000 0.720
S 8,641 0.000 2.760 0.120 0.167 0.028 1.394
SIO2 8,641 0.000 8.740 2.492 0.980 0.972 0.396
Table 4. Model statistics
Valid Min Max Mean Standard
Desviation
Variance Coefficient of
Variation
FEM 30,538 0.000 60.110 19.446 9.897 97.944 0.509
FET 30,306 57.540 70.610 67.449 1.564 2.445 0.023
Al2O3 16,402 0.000 3.072 0.881 0.378 0.143 0.429
CAO 30,306 0.363 4.699 0.795 0.296 0.088 0.372
MGO 16,402 0.000 2.103 0.352 0.205 0.042 0.581
P 30,306 0.000 0.114 0.017 0.011 0.001 0.642
S 30,306 0.000 2.287 0.131 0.152 0.023 1.162
SIO2 30,306 0.926 7.714 2.599 0.905 0.819 0.348
Figure 8. Histogram of the 2016 database (EDA
2016 in red) vs. the database after improvement in 2023
(Implemented 2023 in green)
Figure 9. Graph of tonnage-curve vs. grade (EDA 2016 in
green) compared to the database after improvement in 2023
(Implemented 2023 in blue)