Artificial intelligence is a key tool to transform the “tsunami of data” being generated in today’s mines into vital orebody knowledge that can feed better downstream decisions, Imdex chief geoscientist Dave Lawie has told the World Mining Congress in Australia.
Lawie, a veteran of more than four decades in the mining and energy industries, said AI-generated analysis and visualisation were the “last pieces of the puzzle”.
Chairing a session on AI and rapid resource modelling, and in a separate presentation, Lawie said the “Internet of Geosensing” – networks of cloud-connected tools and sensors gathering information across mine sites – delivered plenty of valuable data.
“But the last thing you can do with a mine is generate a lot of raw data because no-one has time to look at it,” he said at the Brisbane conference and exhibition.
“This is where AI comes in.
“If we go in with an Internet of Geosensing approach and start collecting data, we need to use AI to turn it into information which is directly usable by the mining operation in virtually real time.
“If you can’t do that no-one will use the data; it’s unreasonable to expect people to.”
Lawie said a general lack of training data was currently a roadblock to successful use of AI in mining.
“If we are going to get this data and turn it into a spatial prediction we need to know what it is we are trying to predict and we need to know how to train the model to do the predictions so we need the mines to give us the measured outcomes we are trying to predict with this data,” he said.
Orebody knowledge, delivered in real time upstream of the mining process, was an asset from which value could be realised. But, for various reasons, a lot of companies stopped at that point without realising the value, Lawie said.
“Many companies think acquiring the orebody knowledge is the purpose,” he said. However, “it doesn’t get communicated in a useful way, the tools for realising the value are inadequate, there’s a lack of know-how or willingness to apply it within a change-management process [and] the value needs to be articulated financially.”
Speaking ahead of his presentation, Lawie said more mining companies were coming to the realisation that they needed either better orebody knowledge or a better understanding of the data already obtained.
“There are a lot of mining operations around the world now who are encountering more problematic ores at greater depths, lower grades, more variability than they are used to and it’s affecting their ability to maintain profit margins,” he said.
“A lot of the higher-grade deposits of orebodies have been mined out. In times of lower prices, mining companies will cherry pick the highest grades so they can still be profitable but then they are left with lower grades and a mine that is compromised.
“They need orebody knowledge to get at these lower-grade materials and get them out at a good profit.
“It’s a theme that’s running through the industry.
“But they don’t have the data – they’ve never measured it.”
Imdex has developed and is commercialising its Blast Dog blast-hole sensing and physical measurement technology, which is said to provide “near real-time orebody knowledge”. The technology is semi-autonomously deployed for logging material properties and blast-hole characteristics at high spatial density across mine benches and sites.
Lawie said Blast Dog was generating unrivalled pre-blast rock knowledge and a 3D view into the bench that could facilitate improved blast designs and optimise explosive costs.
It also offered ESG benefits through the potential for precision mining and optimisation of downstream processes such as grinding, blending and stockpiling.