Artificial intelligence (AI) has burst onto the public scene in 2023, and behind this craze is a number of companies that have been using AI in various scenarios in recent years.For the mining industry, AI has also become an area that many leading mining companies have explored and tried to develop pilot products, but it has not yet completely changed the way the industry operates.
Ai is still a misunderstood technology, with a common explanation being "a computer as smart as a human".But this is not the AI of 2023, and the people who create AI create them to solve specific problems, not to mimic humans.There are two broad uses for what is currently thought of as AI: making predictions or optimizing existing processes.
01 AI mining
In mining, AI is gradually being applied throughout the value chain.While specific applications vary widely, the common categories of problems found in each area of the value chain are easy to understand.Finding economic resources among thousands and complex survey and drilling results is a prediction problem;We need to predict where the rock mass is likely to be, then test against that prediction, and the results form feedback to improve the next forecast.Dispatching trucks at the mine site is an optimization problem;We want to minimize fleet idle time and maximize production throughout the day, and there are various quantitative and variable factors.
Key areas of AI in the mining value chain in 2023:
Detection (Forecast)
Digging (optimization)
Handling (optimization)
Maintenance (forecast)
Safety (forecast)
ESG (Optimization)
02 Digital Twin
Digital Twins can map certain scenes from the real world into software so that we can analyze and manipulate data and make changes in the real world.How is this different from what we do today?
Today, we have many systems doing different things, many people doing different things, and when something goes wrong, people look inside their isolated systems to gather data about what to do.The problem is that everyone may see different data from the maintenance, control and planning systems, some people look at the problem on site and diagnose it with their senses, while others imagine the problem thousands of kilometres away, and everyone looks at different layers of the problem in a different way.
Digital twinning creates clarity and context by creating a visual picture of the problem, in which all people can view the same data (from many systems) in the same way.This allows control data to be seen by maintenance, engineering data to be seen by planning, and data conflicts to be observed and captured simultaneously, thus improving decision making.Decisions made through digital twinning flow back into the system and create changes in the real world, which are then reflected in the twinning system.
03 Digital twinning in mining
In the mining community, there are 2 main twin categories that reflect the industry:
Asset twinning
Process twinning
Mining as a large asset and process driven industry, both business processes, as well as the physical and chemical processing of ores, are asset driven.Both assets and processes depend on work;Work (e.g. maintenance, planning, upgrades) is done on them and changes them, and both "work" and are changed by that "work".Therefore, any digital twin in the mining industry must deal with asset, process, and work data to create clarity and context.
A semi-autogenous mill twinning system is an example: asset twinning will focus on mechanical, temperature and vibration, power consumption, maintenance of execution and planning, engineering data.Process twinning will focus on grinding efficiency, consumables usage, input and output flows.The digital twin systems may be used for very different purposes, but they are clearly inextricably linked, because in the real world it makes no sense to separate the flow of matter from the flow of progress.
04 Digital twinning and artificial intelligence
What do these two technology categories have in common?They both seek to improve decision-making.Ai optimizes and predicts, and digital twin provides clarity and context.Bringing them together gives decision-makers a clearer picture of the optimization and prediction of assets, processes and jobs.Ai offers the ability to improve certain decisions and learn how to improve them further;Digital twinning allows humans to compare AI predictions and optimizations to the rules-based systems and processes we use every day.
As AI evolves, the ratio of predictions to rules will shift in terms of performance and across use cases, and the application of predictions to violate rules will increase.Humans will need to master the clarity and context of the problem to resolve these conflicts and enable prediction and optimization to occur without negative consequences.Ai and digital twinning are rapidly maturing and proving themselves in a variety of mining scenario-specific applications.
As the ability to predict and optimise specific metrics increases, there is an expectation that these insights will be used to improve overall mine operations.In most cases, this will require close collaboration with functional departments to understand changes to rules-based business processes and the impact of changes on processes, jobs and assets.Ai has the potential to completely change the way we make decisions, but with decision-making power still firmly in human hands for decades to come, our industry needs to combine personal experiences on the ground to find the right fit between machines and humans.The digital twin is a transformational hub that connects people, processes and technology, and manages these transformations in a structured, risk-averse way.
