Value through

Artificial
intelligence
2024

Opportunities to Adopt AI in Copper Mining

 Context

Ai in simple words

AI in the mining industry refers to artificial

intelligence technology that enhances operational
efficiency, safety, and decision-making processes.
Traditional AI in mining involves analyzing large
datasets to predict equipment failures, optimize
resource allocation, and improve overall
performance.
AI technology can process and interpret vast
amounts of data from various sources, providing
insights that enable smarter, data-driven decisions.

This includes predicting maintenance needs,

optimizing operational schedules, and enhancing
safety protocols.
 Context

The AI Journey

2021 2024+
2016 2019
Insight Team Computer Vision
Machine Learning NLP
Managed Analytics Event Detection
Event Prediction Voice Reporting
Automatic Generation of Reports
Automatic Risk Classification
Gen AI
 Context

How AI Fits into Our

Mining 4.0 Model:
Where Are the
Opportunities? Operational value chain

1. OPERATIONAL PROCESSES 2. DIGITAL INTEGRATION 3. CONTEXTUAL CHALLENGES

Solutions that optimize operational processes Solutions that Enable Integrated planning and Solutions that support enterprise-wide
or solve a specific problem Decision-Making business functions and support processes

Selective Extraction Water Management & Environmental Awareness

Intelligent Mining Integrated Operations

People & Process Safety

Layers
Competitive Movement
Situation Awareness
Industrial Optimization Sustainable & Socially responsible growth
Enhanced Value chain Visualization
Connected Operator 4.0 Digital & Analytic Laboratory

Drones & Robots Asset Performance Management Energy Transition & Decarbonization

Technological Foundation

Technical Generative AI & Data Digital Human

AI AV Connectivity Enhanced
Essentials Platform Engineering

Industrial Cybersecurity Computing Vision Drones Bigdata & IoT Telepresence BIM Virtual Reality
Sensors & Automation Large Language Model Robots Digital Twins Collaboration Modular Construction Augmented Reality
Cloud & Computing Virtual Expert Assist Tele-remote Predictive Modeling Telecommunications 3D Scanning Wearables
Autonomous Vehicles Mobile Solutions 3D Printing Exoskeleton

Support value chain

 Context

Challenges in AI Adoption in Mining

Operational Improvement Maintenance Safety

3 experiences 2 Experiences 2 Experiences

Operational improvement in the mining Maintenance in the mining industry Safety in the mining industry involves
industry refers to the process of enhancing involves the processes and strategies implementing measures and strategies
the efficiency and effectiveness of mining implemented to ensure the reliability to protect workers and operations from
operations. This includes reducing and efficiency of equipment and potential hazards. Improving proactive
downtime, maximizing throughput, and operations. Enhancing maintenance decisions, risk planning, and conducting
enabling smarter decision-making. plans, making better decisions, and more efficient root cause analyses are
optimizing budgets are key aspects of essential components of a robust safety
Reducing downtime, maximizing this approach. approach.
throughput, and enabling smarter
decision-making. Enhancing maintenance plans, making Improving proactive decisions, risk
better decisions, and optimizing planning, and conducting more efficient
budgets. root cause analyses.
Operational Improvement

Operational Improvement
Reducing downtime, maximizing throughput, and
enabling smarter decision-making.
 Operational Improvement

Optimizing Mining Operations

through AI
In operational contexts, AI plays a crucial role in optimizing the extraction
and processing of copper

AI models can generate optimal mining plans by simulating various

extraction scenarios and analyzing their outcomes. This includes
determining the most efficient ways to drill, blast, and process ore based
on predictive models that account for equipment performance, weather
conditions, and other variables.

The result is a significant enhancement in

operational efficiency, cost reduction, and
increased yield, all while ensuring compliance with
safety and environmental standards.
 Image Analysis Operational continuity Drilling&Blasting Medium complexity
Powered by Impact Process Affinity MVP
Operational Improvement

Detecting
uncrushables
with AI
Detecting uncrushables is vital for efficiency and
safety in copper mining.

These materials, such as metals or hard rocks, can

damage crushing equipment, leading to costly
repairs and downtime. Generative AI (Gen AI)
identifies uncrushables by analyzing real-time sensor
data.

This reduces risks and optimizes resources,

improving operational continuity and lowering repair
costs. Detection with Gen AI is a key tool for
efficiency and safety in copper mining.

Impact / Estimated Value:

• Over 90% of uncrushables detected

• Zero impact to operational continuity due to
unprogrammed detentions caused by uncrushables in
the crushing process.
Operational Improvement

Detecting uncrushables with AI

 ML models Proccess efficiency Flotation Mid/high complexity
Powered by Impact Process Affinity MVP
Operational Improvement

Plant process A E

optimization
B

The tool is a monitoring and analysis system for the flotation

process in a mineral processing plant. In the "Current Scenario"
section, it provides real-time metrics such as SAG mill throughput
(Sum TPH SAG), particle size (P80), and residence time, which are
crucial for assessing the plant's efficiency.

The "Expected Scenario" section projects future performance,

comparing outcomes with and without implementing
recommendations. This helps users understand potential C
improvements in efficiency or quality by following suggested
changes.

The tool also includes an "Incoming Minerals" section, which

presents data on the mineralogy of materials being processed,
helping anticipate changes in input material. The
"Recommendations" section provides context and reliability
indicators for these suggestions, supporting informed decision- D
making to enhance process performance.

Impact / Estimated Value:

• The value capture is of the order of USD 3.8MM in a

fiscal year ( from client definition).
 ML models Proccess efficiency Flotation Mid/high complexity
Powered by Impact Process Affinity MVP
Operational Improvement

Plant process
optimization

“In the field we know what we have to do, we have our plans
(weekly, daily). What we do not know, is what kind of
material/mineral is going to arrive.
- Plant Coordinator

A tool like this can bring consistency and stability to the

Rougher operation, a lot of live-testing could be avoided by
having recipes for different minerals.
- Control Supervisor
 ML models Decision making End to end Medium complexity
Powered by Impact Process Affinity MVP
Operational Improvement

Mineral tracking

This is a tool that allows tracking of transported

material in near real time, using data from systems
that store mineralogical data from the mine, as well
as shovel and truck data from the loading and
transport systems.

The tool estimates how the material is divided on

the "conveyor belts" using information from sensors
(weight meters). Finally, the estimated information is
compared with the plan information and the
percentage of compliance with the plan can be
measured.

In this way, decision making is supported by

delivering the information on time.

Impact / Estimated Value:

• The level of compliance with the planned is achieved in

a time close to real time.
 Maintenance

Maintenance
Enhancing maintenance plans, making better
decisions, and optimizing budgets.
 Maintenance

Improving Maintenance Plans: AI analyzes historical and real-time

data to predict equipment failures, enabling more accurate and
Enhancing Maintenance with AI in Mining
timely maintenance scheduling.

Better Decision-Making: AI-driven insights offer precise, data-

backed recommendations, aiding in informed decisions regarding
repairs and replacements.
Artificial Intelligence (AI) is transforming maintenance practices in
the mining industry by providing predictive insights and Optimizing Budgets: Predictive maintenance reduces unexpected
optimizing resources. Leveraging AI technology, mining breakdowns, allowing for better budget allocation and minimizing
operations can achieve higher efficiency and cost-effectiveness. unnecessary expenses, thus extending equipment lifespan.
 Maintenance

Update of Maintenance Strategy Leveraging Artificial Intelligence

Situation
Our client, a traditional thermal generator The conventional plants are experiencing
with assets installed since the 1970s, is facing constant changing demand, which requires
a changing scenario due to the introduction them to change their maintenance strategies
of renewable energies into the national power more frequently than what was originally
grid. designed for those plants.
Pilot Scope
• Industry: Power Generation • Components not covered in scope:
• Country: Chile o Valves
• Equipment type: Vertical Pump o Measurement instruments
• Equipment brand/model: KSB WKTA 200/4 o Piping
• Components covered in scope:
o Pump
o Electric Motor
Project Journey
• Key milestones: • Output from equipment strategies
o Kick-off (seting the expectations and development:
scope) o 20 Failure Modes
o First draft equipment strategies review o 20 Primary Tasks
with the client o 17 Secondary Tasks
o Deliverable handover (loadhseet) o 3 PM and 17 CM task types
o Client feedback o 39 Equipment strategy fields completed
at 100% Impact / Estimated Value:
o 11 components identified and analyzed
• 80% Higher accuracy than client current maturity for
this specific asset
• 60% Faster than traditional approach
 LML Efficiency on tasks Support areas Low complexity
Powered by Impact Process Affinity MVP
Maintenance

Gartner study says that workers spend 20%-30%

Knowledge management of their time searching and interpreting

information (1.6h to 2.4h a day)1.
involves organizing and This estimation assumes that are trained
presenting information in a workers and the data sources are not new to
them.
clear and trustworthy
manner.
What do I have to do if the radio fails in
virtual engineer can streamline the process of Question the Integrated Operations Center?
troubleshooting equipment issues in an Integrated
Operations Center
Data Sources
1.Data Sources: It accesses hundreds of unstructured
Hundreds of unstructured
documents, each containing hundreds of pages.
documents with hundreds of
pages each
2.Filter and Find Relevant Documents: The virtual
engineer filters through these documents to identify
and extract the most relevant information.
Filter and find relevant
3.Interpret Information: It interprets the extracted documents
information to provide a comprehensive understanding
of the issue.

4.Answer: Finally, it provides a clear and actionable Interpret information

answer, such as "Initiate the Z8 procedure where the
site supervisor takes over the operation." Initiate the Z8 procedure where the site
Answer Impact / Estimated Value:
supervisor takes over the operation
• Over 30% improvement in efficiency by providing end-
users with critical information and contextual variables.
 Safety

Safety
Improving proactive decisions, risk planning, and
conducting more efficient root cause analyses.
 Safety

Enhancing Mining Safety with Generative AI

Generative AI offers transformative capabilities in enhancing

safety within the copper mining industry.

By leveraging real-time data from a network of

sensors throughout mining operations, AI
models can predict and simulate potential
hazards, from equipment malfunctions to
structural failures.
This predictive ability allows for preemptive actions to prevent
accidents, ensuring the safety of the workforce
 Image Analysis Volume of accidents Plant processes Low complexity
Powered by Impact Process Affinity MVP
Safety

Ai-Eyes on
safety
Eyes on Safety can monitor existing CCTV
cameras online and detect patterns and trends
that impact occupational risk in industrial
production areas. Customers using the solution
have been able to avoid accidents and critical
situations by defining more proactive occupational
safety approaches.

It uses a neural network trained to identify the use

of specific protective gear, such as helmets, ear
protection, goggles, half-face and full-face respirator
filters, and surgical masks. I

t analyzes video signals from CCTV cameras in the

client's network, processed by the software, which
generates alerts sent to a reporting interface. This
system enhances safety and regulatory compliance,
reducing risks and optimizing PPE management in
the industry.
Impact / Estimated Value:

• Reduce disruption of Operations: Injuries and accidents,

• Reduce Costs: Medical expenses, compensation
• Workforce Morale: Frequent accidents negatively impact
employee morale
 LLM models Time on investigation Mine+Plant Low complexity
Powered by Impact Process Affinity MVP
Safety

Expert Advisor for

Accident Investigation

1. Time and Specialist Requirement

Digital Assistant that supports the
entire Accident Investigation
process, managing its complexity Impact: The future state aims to streamline the
(effort, cost, time, technical investigation process, reducing the dependency on
specialization). specialists and accelerating the timeline.

Depending on the context and 2. Structure of Investigation Data

type of accident, it suggests how
to conduct the investigation Impact: This transformation improves data accessibility
according to the required method, and usability, facilitating better-informed decisions and
recommends questions and quicker responses.
elements to verify, provides an
inventory of relevant controls and
defenses, and offers libraries of 3. Basis of Investigation
precursors. It helps improve Impact: Shifting to data-driven investigations enhances
productivity, quality, and avoid objectivity and reliability, leading to more accurate
biases. conclusions and recommendations.

4. Data Processing

•Impact: Automation of data processing not only saves

time and effort but also improves the accuracy and Impact / Estimated Value:
consistency of data handling.
• Faster and better-conducted investigations, increasing
the quantity and quality of data. Opportunity for time
savings in investigations: 43%.
 LLM models Time on investigation Mine+Plant Low complexity
Powered by Impact Process Affinity MVP
Safety

Expert Advisor for

Accident Investigation

Impact / Estimated Value:

• Faster and better-conducted investigations, increasing

the quantity and quality of data. Opportunity for time
savings in investigations: 43%.
 ML Models Volume of accidents Plant processes Low complexity
Powered by Impact Process Affinity MVP
Safety

Building blocks on safety

Critical Risk Analysis Prediction at the Risk level

An AI engine learns how experienced HSE engineers classify critical Forecast at the risk level for the next week with an alert for critical events,
events and is able to apply this to all new safety reporting accidents with people, property accidents.
contributions, detecting severity/criticality and what factor the risk
is associated with: equipment, tasks, behavior, circumstances, etc.
 Input

Your Voice
Context

“The real challenge in integrating AI in mining lies

not in the technology itself, but in earning the trust
of those who will use it every day”
 2024

Opportunities to Adopt AI in Copper Mining