XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3 1081
phases, such as in the design and operations of mines, rely
on accurate resource definition. With increased automa-
tion, the design of process flowsheets can be streamlined
by leveraging AI. Likewise, systems diagnostics can be
improved in equipment monitoring by utilizing advanced
machine vision and data management systems.
Connecting all these simplified systems requires care-
ful planning and well-defined objectives (Figure 12).
Generative AI can connect simplified systems with clear
objectives developing such systems will require connect-
ing the resource definition data (that is, carrying informa-
tion on metal grades and compositions) across the mine,
processing plant, and mining equipment. This step requires
extensive data tracking and the generation of new data
points to increase the resolution.
Data resolution of the resource block can be increased
by utilizing a combination of integrated model develop-
ment, model fitting and sensors, Figure 13. AI enabled data
analysis can be used to fill-in the gaps in the analysis tech-
niques used to predict the metallurgical outputs from an
operating mine.
CONCLUSIONS
The short answer to the question, ‘Can AI save the min-
ing industry?’ is NO. AI software requires quality data to
train neural networks and machine-learning algorithms. In
operating mines, data is sourced from functioning instru-
mentation and sensors in process equipment and any
machinery part of the mining ecosystem from non-instru-
mented data sources, such as resource definition from drill
hole samples, metallurgical test work, and user experience
information from human-machine interactions.
An effective AI system installed to predict future events
or performances requires information from WORKING
instrumentation/sensors and QUALITY information that
can be time-coded and tracked for the entire value chain.
Working instrumentation and quality data are critical to
train the algorithms to make correct predictions of the
working system, which, ultimately, requires human inter-
actions to make sense of the data in the context of the
mining operation. So, instead of addressing the question,
‘Can AI save the mining industry?’, the focus of this paper
addresses, ‘How can AI be used as an effective solution for
the mining industry?’
The key to developing an effective AI system for the
mining industry is by:
• Building robust sensors and instrumentation systems
that can minimise faults in the measurement and
increase measurement accuracy
• Leveraging ore characterisation data measured
in realtime to benchmark calibrate sensors and
instrumentation
Figure 12. Building connected systems with AI (Chandramohan, 2023)
phases, such as in the design and operations of mines, rely
on accurate resource definition. With increased automa-
tion, the design of process flowsheets can be streamlined
by leveraging AI. Likewise, systems diagnostics can be
improved in equipment monitoring by utilizing advanced
machine vision and data management systems.
Connecting all these simplified systems requires care-
ful planning and well-defined objectives (Figure 12).
Generative AI can connect simplified systems with clear
objectives developing such systems will require connect-
ing the resource definition data (that is, carrying informa-
tion on metal grades and compositions) across the mine,
processing plant, and mining equipment. This step requires
extensive data tracking and the generation of new data
points to increase the resolution.
Data resolution of the resource block can be increased
by utilizing a combination of integrated model develop-
ment, model fitting and sensors, Figure 13. AI enabled data
analysis can be used to fill-in the gaps in the analysis tech-
niques used to predict the metallurgical outputs from an
operating mine.
CONCLUSIONS
The short answer to the question, ‘Can AI save the min-
ing industry?’ is NO. AI software requires quality data to
train neural networks and machine-learning algorithms. In
operating mines, data is sourced from functioning instru-
mentation and sensors in process equipment and any
machinery part of the mining ecosystem from non-instru-
mented data sources, such as resource definition from drill
hole samples, metallurgical test work, and user experience
information from human-machine interactions.
An effective AI system installed to predict future events
or performances requires information from WORKING
instrumentation/sensors and QUALITY information that
can be time-coded and tracked for the entire value chain.
Working instrumentation and quality data are critical to
train the algorithms to make correct predictions of the
working system, which, ultimately, requires human inter-
actions to make sense of the data in the context of the
mining operation. So, instead of addressing the question,
‘Can AI save the mining industry?’, the focus of this paper
addresses, ‘How can AI be used as an effective solution for
the mining industry?’
The key to developing an effective AI system for the
mining industry is by:
• Building robust sensors and instrumentation systems
that can minimise faults in the measurement and
increase measurement accuracy
• Leveraging ore characterisation data measured
in realtime to benchmark calibrate sensors and
instrumentation
Figure 12. Building connected systems with AI (Chandramohan, 2023)