Vale to use artificial intelligence at its nickel mine
Mining company, Vale, plans to use machine-learning and artificial intelligence to explore for new areas to drill. Idha Valeur reports.
Brazilian mining company, Vale, has engaged GoldSpot Discoveries to use machine-learning (ML) tools at its Coleman nickel mine in Ontario, Canada.
Vale Manager, Corporate Affairs and Sustainability, North Atlantic Operations and Asian Refineries, Angie Robson, told Materials World GoldSpot uses artificial intelligence (AI) and ML to analyse new and older data to recognise trends, as well as identify information that previously went under the radar. ‘It is very difficult for humans to take in every facet of information gathered at a mine every day – over 50 years – and think about it all at once to analyse,’ she said.
‘It’s just not the way humans are built. With ML and AI, it can be done when guided by geological experts. So all that data, from 50 years ago to present, can be leveraged to produce potential new exploration theories to be vetted by site geologists for exploration potential.’
Although digital technologies such as ML and AI can drive improvements in the mining industry, Robson highlighted that it is mostly a tool to support mineral exploration and optimising resources. However, it’s not about comparing technologies with established ones, it is about having a new tool in the box.
‘Think of times when geophysics added new exploration tools. It never replaced traditional exploration techniques, but rather complemented them and gave another layer of risk reduction for exploration targeting,’ she said. ‘This is what AI and ML are doing now for the industry – not replacing tried and true methods but rather analysing data in a novel way to reduce risk and increase confidence in drill targets.’
Saving cost and time
As well as the ability to reduce operational risk, according to Robson, AI and ML can reduce exploration costs and increase the efficiency of the exploration phase of looking for new ore. ‘ML provides another layer of impartial data analysis to give us more confidence in drilling and exploring in some areas, while also helping to reduce the overall search space. ML and AI provide ways of interrogating data, seeing if your exploration model is backed by the data, and generating other ways of exploring based on unconventional factors that may be more efficient. GoldSpot believes this added layer of data analysis and vetting helps minimise exploration risk overall,’ she said.
Further, Robson explained that Vale is working with GoldSpot to accelerate the near-mine exploration programme at Coleman mine in 2020. ‘Ultimately, this work will produce new drill targets for our exploration team to test next year. Some of these newly identified targets will lead to new discoveries and orebodies that will extend the life of Coleman mine and potentially keep exploration active for years to come.’