7
the correct approach if there is a continual review of how
value can be created for a project. As demonstrated by our
analysis, even minor deviations in throughput can lead to
disproportionate impacts on NPV.
To address these challenges, we advocate for an inte-
grated geometallurgical approach that increases both the
quantity and variety of comminution tests during the early
stages of project development. This does not imply that the
ratio of comminution tests to assay tests should approach
unity but does emphasize the importance of early integra-
tion of sampling i.e., identifying patterns in the geology,
mineralogy and elemental composition and testing the com-
minution behaviour of identified classes of ore. Artificial
Intelligence (AI) models, such as Deep Neural Networks
(DNN) and Particle Swarm Optimization (PSO), serve as
powerful regression tools capable of estimating comminu-
tion parameters based on geochemical and mineralogical
data (De Almeida et al., 2024a). These models are trained
using a comprehensive dataset of comminution test results,
requiring enough tests to ensure accuracy and reliability.
Once trained and validated for a deposit, the models can
predict comminution parameters for samples where only
geochemical and mineralogical data are available. Notably,
these models can also be developed in the absence of miner-
alogical data, using only geochemistry as input, with a mar-
ginal trade-off in prediction accuracy for the comminution
data output (De Almeida et al., 2024b). By understand-
ing ore hardness variability more comprehensively across
the orebody, mining companies can achieve more reliable
throughput forecasts, enabling better alignment between
ore block characteristics, operational performance and
project revenue. This, in turn, reduces financial risks and
strengthens confidence in project outcomes.
Ultimately, a more balanced and rigorous approach
to comminution testing supports robust circuit design
and ensures the economic viability of mining projects. By
bridging the current gap in sampling practices, the industry
can enhance its ability to deliver predictable and sustainable
performance in increasingly complex and capital-intensive
operations.
REFERENCES
[1] Angove, J. E, &Dunne, R. C. (1997). A Review
of Standard Physical Ore Property Determinations.
World Gold ‘97 Conference, Singapore.
[2] Bond, F.C. (1952). “The Third Theory of
Comminution.” Transactions of the AIME, Vol. 193,
pp. 484–494.
[3] Bueno, M., Foggiatto, B., and Lane, G. (2015).
Geometallurgy Applied in Comminution to
Minimize Design Risks. SAG 2015 Conference,
Sixth International Conference on Semiautogenous
and High-Pressure Grinding Technology, Vancouver,
20–24 September 2015.
[4] Bueno, M., Torvela, J., Chandramohan, R., Matus,
T.C., Liedes and T., Powell, M.S. (2021). The
double wheel breakage test, Minerals Engineering,
Volume 168, 106905, ISSN 0892-6875, https://doi
.org/10.1016/j.mineng.2021.106905.
[5] C, Bailey., G, Lane., Morrell, S., and Staples, P.
(2009). What Can Go Wrong in Comminution
Circuit Design? Proceedings of the Tenth Mill
Operators’ Conference.
[6] De Almeida, T., Nicolau, A., Schirru, R. and Bueno,
M., (2024). Development of a deep neural network
and a PSO algorithm to predict ore hardness using
X-ray diffraction and atomic emission spectroscopy.
Minerals Engineering, vol 213.
[7] De Almeida, T., Nicolau, A., Schirru, R. and Bueno,
M. (2024). Nuclear Analytical Techniques Applied to
Determine Ore Hardness Using Artificial Intelligence.
International Nuclear Atlantic Conference. Rio de
Janeiro, 2024.
[8] Doll, A. (2018). Designing an optimal comminution
sampling program for geometallurgy. Procemin 2018
– 14th International Mineral Processing Seminar
(pp. 415–421). Santiago: Gecamin.
[9] Couët, F., Goudreau, S., Makni, S., Brissette, M.,
Brissette, H., Longuépée, G., Gagnon and Rochefort,
C. (2015) A New Methodology for Geometallurgical
Mapping of Ore Hardness. In Proceedings SAG 2015.
Sixth International Conference on Semiautogenous
and High-Pressure Grinding Technology, 20–24
September 2015.
[10] Koch, G. and Link, R. (1971). The coefficient of varia-
tion a guide to the sampling of ore deposits. Economic
Geology 1971 66 (2): 293–301. https://doi
.org/10.2113/gsecongeo.66.2.293.
[11] Kojovic, T. (2016). HIT – a Portable Field Device for
Rapid Testing at Site. Proceedings of the thirteenth
Mill Operators’ Conference.
[12] Lane, G. (2010). What is a competent ore and how
does this effect comminution circuit design? In
Kracht, W., Kuyvenhoven, R., Montes-Atenas, G.
(Eds.), Procemin 2010 – 7th International Mineral
Processing Seminar (pp. 33–44). Santiago: Gecamin.
[13] Lane, G., Foggiatto, B., Braun, R., Bueno, M.,
Staples, P., and Rivard, S. Comminution Circuit
Design Considerations, in Muinonen J, Cameron R
and Zinck J (Editors) 49th Annual Canadian Minerals
the correct approach if there is a continual review of how
value can be created for a project. As demonstrated by our
analysis, even minor deviations in throughput can lead to
disproportionate impacts on NPV.
To address these challenges, we advocate for an inte-
grated geometallurgical approach that increases both the
quantity and variety of comminution tests during the early
stages of project development. This does not imply that the
ratio of comminution tests to assay tests should approach
unity but does emphasize the importance of early integra-
tion of sampling i.e., identifying patterns in the geology,
mineralogy and elemental composition and testing the com-
minution behaviour of identified classes of ore. Artificial
Intelligence (AI) models, such as Deep Neural Networks
(DNN) and Particle Swarm Optimization (PSO), serve as
powerful regression tools capable of estimating comminu-
tion parameters based on geochemical and mineralogical
data (De Almeida et al., 2024a). These models are trained
using a comprehensive dataset of comminution test results,
requiring enough tests to ensure accuracy and reliability.
Once trained and validated for a deposit, the models can
predict comminution parameters for samples where only
geochemical and mineralogical data are available. Notably,
these models can also be developed in the absence of miner-
alogical data, using only geochemistry as input, with a mar-
ginal trade-off in prediction accuracy for the comminution
data output (De Almeida et al., 2024b). By understand-
ing ore hardness variability more comprehensively across
the orebody, mining companies can achieve more reliable
throughput forecasts, enabling better alignment between
ore block characteristics, operational performance and
project revenue. This, in turn, reduces financial risks and
strengthens confidence in project outcomes.
Ultimately, a more balanced and rigorous approach
to comminution testing supports robust circuit design
and ensures the economic viability of mining projects. By
bridging the current gap in sampling practices, the industry
can enhance its ability to deliver predictable and sustainable
performance in increasingly complex and capital-intensive
operations.
REFERENCES
[1] Angove, J. E, &Dunne, R. C. (1997). A Review
of Standard Physical Ore Property Determinations.
World Gold ‘97 Conference, Singapore.
[2] Bond, F.C. (1952). “The Third Theory of
Comminution.” Transactions of the AIME, Vol. 193,
pp. 484–494.
[3] Bueno, M., Foggiatto, B., and Lane, G. (2015).
Geometallurgy Applied in Comminution to
Minimize Design Risks. SAG 2015 Conference,
Sixth International Conference on Semiautogenous
and High-Pressure Grinding Technology, Vancouver,
20–24 September 2015.
[4] Bueno, M., Torvela, J., Chandramohan, R., Matus,
T.C., Liedes and T., Powell, M.S. (2021). The
double wheel breakage test, Minerals Engineering,
Volume 168, 106905, ISSN 0892-6875, https://doi
.org/10.1016/j.mineng.2021.106905.
[5] C, Bailey., G, Lane., Morrell, S., and Staples, P.
(2009). What Can Go Wrong in Comminution
Circuit Design? Proceedings of the Tenth Mill
Operators’ Conference.
[6] De Almeida, T., Nicolau, A., Schirru, R. and Bueno,
M., (2024). Development of a deep neural network
and a PSO algorithm to predict ore hardness using
X-ray diffraction and atomic emission spectroscopy.
Minerals Engineering, vol 213.
[7] De Almeida, T., Nicolau, A., Schirru, R. and Bueno,
M. (2024). Nuclear Analytical Techniques Applied to
Determine Ore Hardness Using Artificial Intelligence.
International Nuclear Atlantic Conference. Rio de
Janeiro, 2024.
[8] Doll, A. (2018). Designing an optimal comminution
sampling program for geometallurgy. Procemin 2018
– 14th International Mineral Processing Seminar
(pp. 415–421). Santiago: Gecamin.
[9] Couët, F., Goudreau, S., Makni, S., Brissette, M.,
Brissette, H., Longuépée, G., Gagnon and Rochefort,
C. (2015) A New Methodology for Geometallurgical
Mapping of Ore Hardness. In Proceedings SAG 2015.
Sixth International Conference on Semiautogenous
and High-Pressure Grinding Technology, 20–24
September 2015.
[10] Koch, G. and Link, R. (1971). The coefficient of varia-
tion a guide to the sampling of ore deposits. Economic
Geology 1971 66 (2): 293–301. https://doi
.org/10.2113/gsecongeo.66.2.293.
[11] Kojovic, T. (2016). HIT – a Portable Field Device for
Rapid Testing at Site. Proceedings of the thirteenth
Mill Operators’ Conference.
[12] Lane, G. (2010). What is a competent ore and how
does this effect comminution circuit design? In
Kracht, W., Kuyvenhoven, R., Montes-Atenas, G.
(Eds.), Procemin 2010 – 7th International Mineral
Processing Seminar (pp. 33–44). Santiago: Gecamin.
[13] Lane, G., Foggiatto, B., Braun, R., Bueno, M.,
Staples, P., and Rivard, S. Comminution Circuit
Design Considerations, in Muinonen J, Cameron R
and Zinck J (Editors) 49th Annual Canadian Minerals