1402 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
performance in a more holistic and concrete manner, using
actual operational data rather than proxies such as metal-
lurgical laboratory samples. This also requires the use of
modeling techniques capable of handling the time-based
complexity of the system. For this, discrete event simula-
tion (DES) is a well-established methodology for modeling
large and complex systems, especially those with non-linear
behavior.
It is important to note the DES models are typically
developed to investigate strategic concerns, rather than
to track real-time system behavior in granular detail. The
model developed in this study is intended for use in inves-
tigating the role of geometallurgical attributes on process
performance on a macro-scale, rather than to be used as a
short interval control model. Potential outcomes may indi-
cate a benefit from changing ore routing criteria for example
or suggest ways in which tweaks the usual operating con-
ditions might result in improved throughput, recovery, or
reduced maintenance, for example. Any proposed changes
would still require careful consideration, including discus-
sion of the underlying ore properties and the causal con-
nection to process performance, prior to implementation.
Literature Review
Geometallurgy is a rather new field within the mining
industry as the term has only been in common use since
the 1990s (Dean, 2019). Research that explicitly links min-
ing and processing operational data also requires a work-
ing knowledge of modern data analytics and management
information systems and techniques, including techniques
that can account for the time delays inherent to the sys-
tem and the volume of data available. Given these inter-
disciplinary considerations, the existing literature related
to DES models applied to geometallurgical use cases and
using observed production data is rather limited.
A handful of studies incorporating both geostatistical
techniques with discrete event simulation models have been
published recently, however. Wilson et al present a case
study of a DES model that combined geostatistical variabil-
ity of tailings dam material with a simulation modeling the
potential operational risks of remining the tails for the pur-
poses of cement production (2021). Saldaña et al use DES
to model a synthetic heap-leaching process with a focus on
improving recovery and efficiency (2019). Another series
of studies use DES to construct a digital twin for the pur-
poses of improving gold production at a mine in the Alhué
district of Chile (Órdenes et al., 2022 Peña-Graf, 2022
Órdenes et al., 2021). On a slightly larger scale, Wilson et
al discuss a regional framework for mine-to-smelter inte-
gration of Chilean copper smelters (2022). A related study
uses DES techniques to support modernization of legacy
copper smelters using advanced process control techniques
(Navarra et al., 2020). Another study addresses the regional
development of refractory gold processing systems (Wilson
et al, 2022). Wilson has also published a study using DES
techniques applied to a digital twin for a Canadian oil
sands operation, with a specific emphasis on the geological
uncertainty related to the quality and heterogeneity of the
bitumen (2021).
Other studies use advanced analytics approaches to
tackle similar research goals of improving the geostatistical
models for a particular ore deposit by directly tying the mod-
eled geologic parameters to production. In a dissertation by
Wambuke, the author uses online production data to gen-
erate real-time work index updates to a geostatistical model
at the Tropicana gold mine (2018). Similarly, Parian (2017)
relies on combining liberation and bulk and size fraction
data to model modal mineralogies in an iron ore processing
circuit to predict mineral processing performance. Koch
presents a recommendation system to optimize ore rout-
ing (2019). Lastly, Tijsseling et al develop a flotation model
for a copper-cobalt deposit in the Democratic Republic of
the Congo that quantifies the influence of mineralogical
variability on process performance (2020). Taken together,
such studies represent first steps toward development of a
more integrated approach to monitoring and reconciliation
of process performance.
Study Overview
In this study, the authors utilize mining and production
data from an active copper porphyry operation in Arizona.
Most of the ore at the operation is oxide, as the deposit
has undergone extensive supergene enrichment. However,
at the deeper levels of the deposit hypogene sulfide ore
remains and is a minor and increasing component of the
production. Sulfide ore is processed via a flotation circuit,
which includes facilities to crush, grind, and float the ore,
resulting in a copper concentrate which is shipped off-site
for further refining. Each shovel load is tracked bucket by
bucket, resulting in a a direct record how many buckets
of ore are required to fill each truck load. From there, the
dump destination and time of dumping for each truckload
is also recorded.
While DES and advanced analytics techniques can be
used to address production throughout the entire process
system from primary crushing to the concentrate pro-
duction, this study focuses on only the first component
of that system. The model described here spans from the
point at which mill ore is dumped into the primary crusher
through a series of conveyors where ore is deposited into
performance in a more holistic and concrete manner, using
actual operational data rather than proxies such as metal-
lurgical laboratory samples. This also requires the use of
modeling techniques capable of handling the time-based
complexity of the system. For this, discrete event simula-
tion (DES) is a well-established methodology for modeling
large and complex systems, especially those with non-linear
behavior.
It is important to note the DES models are typically
developed to investigate strategic concerns, rather than
to track real-time system behavior in granular detail. The
model developed in this study is intended for use in inves-
tigating the role of geometallurgical attributes on process
performance on a macro-scale, rather than to be used as a
short interval control model. Potential outcomes may indi-
cate a benefit from changing ore routing criteria for example
or suggest ways in which tweaks the usual operating con-
ditions might result in improved throughput, recovery, or
reduced maintenance, for example. Any proposed changes
would still require careful consideration, including discus-
sion of the underlying ore properties and the causal con-
nection to process performance, prior to implementation.
Literature Review
Geometallurgy is a rather new field within the mining
industry as the term has only been in common use since
the 1990s (Dean, 2019). Research that explicitly links min-
ing and processing operational data also requires a work-
ing knowledge of modern data analytics and management
information systems and techniques, including techniques
that can account for the time delays inherent to the sys-
tem and the volume of data available. Given these inter-
disciplinary considerations, the existing literature related
to DES models applied to geometallurgical use cases and
using observed production data is rather limited.
A handful of studies incorporating both geostatistical
techniques with discrete event simulation models have been
published recently, however. Wilson et al present a case
study of a DES model that combined geostatistical variabil-
ity of tailings dam material with a simulation modeling the
potential operational risks of remining the tails for the pur-
poses of cement production (2021). Saldaña et al use DES
to model a synthetic heap-leaching process with a focus on
improving recovery and efficiency (2019). Another series
of studies use DES to construct a digital twin for the pur-
poses of improving gold production at a mine in the Alhué
district of Chile (Órdenes et al., 2022 Peña-Graf, 2022
Órdenes et al., 2021). On a slightly larger scale, Wilson et
al discuss a regional framework for mine-to-smelter inte-
gration of Chilean copper smelters (2022). A related study
uses DES techniques to support modernization of legacy
copper smelters using advanced process control techniques
(Navarra et al., 2020). Another study addresses the regional
development of refractory gold processing systems (Wilson
et al, 2022). Wilson has also published a study using DES
techniques applied to a digital twin for a Canadian oil
sands operation, with a specific emphasis on the geological
uncertainty related to the quality and heterogeneity of the
bitumen (2021).
Other studies use advanced analytics approaches to
tackle similar research goals of improving the geostatistical
models for a particular ore deposit by directly tying the mod-
eled geologic parameters to production. In a dissertation by
Wambuke, the author uses online production data to gen-
erate real-time work index updates to a geostatistical model
at the Tropicana gold mine (2018). Similarly, Parian (2017)
relies on combining liberation and bulk and size fraction
data to model modal mineralogies in an iron ore processing
circuit to predict mineral processing performance. Koch
presents a recommendation system to optimize ore rout-
ing (2019). Lastly, Tijsseling et al develop a flotation model
for a copper-cobalt deposit in the Democratic Republic of
the Congo that quantifies the influence of mineralogical
variability on process performance (2020). Taken together,
such studies represent first steps toward development of a
more integrated approach to monitoring and reconciliation
of process performance.
Study Overview
In this study, the authors utilize mining and production
data from an active copper porphyry operation in Arizona.
Most of the ore at the operation is oxide, as the deposit
has undergone extensive supergene enrichment. However,
at the deeper levels of the deposit hypogene sulfide ore
remains and is a minor and increasing component of the
production. Sulfide ore is processed via a flotation circuit,
which includes facilities to crush, grind, and float the ore,
resulting in a copper concentrate which is shipped off-site
for further refining. Each shovel load is tracked bucket by
bucket, resulting in a a direct record how many buckets
of ore are required to fill each truck load. From there, the
dump destination and time of dumping for each truckload
is also recorded.
While DES and advanced analytics techniques can be
used to address production throughout the entire process
system from primary crushing to the concentrate pro-
duction, this study focuses on only the first component
of that system. The model described here spans from the
point at which mill ore is dumped into the primary crusher
through a series of conveyors where ore is deposited into