Artificial Intelligence Bulletin – A.I. in Mining

In this week’s Artificial Intelligence Bulletin – A.I. in Mining we take a deeper dive into how the mining industry is leading the way towards automation and artificial intelligence. How is it already being deployed in mines? What are some of the advantages of A.I. over humans? Read on…
The mining industry is a complex beast that begins with a small group of mineral exploration geologists or even a solo prospector and ends up as raw material in nearly every other area of our economy.
Modern mining is rife with challenges and hazards including; the inability to recruit talented labour due to remote locations and inclement weather, safety issues & injuries arising from close proximity of workers to heavy machinery and energy cost efficiency – to name a few.
It’s not surprising then that leaders in this industry are more consistently turning to artificial intelligence to help solve many of these generations-old problems.
It may seem counter-intuitive that a sand-blasted crater of rock would be a haven for machine learning, but upon closer inspection mines can offer the ideal set of conditions for A.I. to thrive. The amount of real-time, raw & high volume data acquired throughout all stages of the mining life-cycle, from exploration and extraction to processing and production is immense.

Mining automation has swept across the industry with early adopters like Rio Tinto using autonomous 350 tonne haul trucks as early as 2008. Being a pioneer of this new technology paid off as they have experienced not only increased efficiency but also lowered fuel costs and improved worker safety.
Their use of remote controlled and autonomous vehicles helped pave the way for an automation revolution within the industry. What initially began with trucks, loaders and drillers has now swept through all stages of the mining life cycle and inspired a sub-industry of data-analytics and A.I. startups catering exclusively to mining.
Professors Ajay Agrawal, Joshua Gans and Avi Goldfarb of Rotman School of Management, and the authors of the recent AI book, Prediction Machines: The Simple Economics of Artificial Intelligence speak about how mining automation, specifically by Rio Tinto in Western Australia’s isolated Pilbara region, has provided a solution to one of the industries biggest costs: downtime.
Even though most mining operations currently utilize a 24-hour shift work clock, humans are still inhibited by our circadian rhythm and work far less productively at night. They concur that, “mining is the perfect opportunity for full automation precisely because it has already removed humans from so many activities.”

“The real value of this [A.I.] is that it allows for real time optimization of the mine plan when you input drilling data or tweak mining models – AI applications automatically update all other relevant areas and are continually applying machine learning to optimize these results and maximize the amount of minerals we get from mines. It will be a huge factor in reducing costs.” – Goldcorp VP Luis Canepari

A.I. now is swiftly merging itself with all mining stages. Geologists in exploration are using machine learning to review geological data and predict best spots for drilling locations. Falco Resources uses, and has had recent success, with pattern recognition algorithms (known as CARDS – Computer Aided Resource Detection System) to predict and find gold.
Machine maintenance is also a major area of mining where predictive intelligence can have an impact on cost savings and efficiency. Algorithms can learn to identify mechanical pre-failure signs and signatures by assessing real-time equipment sensor data and external influencers such as weather. Likewise, In rock blasting, machine learning has helped better predict fragmentation and aided in explosives selection by analyzing drill hole patterns, geology and myriad other factors.
With almost every area and step in the mining process allowing for improvements through automation it’s only time before each step is fully integrated – aided by machine learning and AI. Indeed Rio Tinto is ahead of the curve again here, building a fully integrated intelligent mine for $2.2 billion. Equipped with advanced robotics, and driverless trucks and trains, the Koodaideri mine serves to solve the islands of automation problem.

This fully automated mining future has some great potential benefits, not only in cost-savings and profitability for the corporations involved but also for sustainability and employment. Changing the way mines are built from the ground up with a focus on electrification, automation and digitization will reduce some of the current burden on the environment.
Automation will also shift the current employee profile to one of a skilled technologist, taking a hands-off approach to handling heavy equipment and improving safety by keeping humans out of harms way. For those labourers still in the field machine learning can assist as well. In a recent experiment, a Congolese copper mine leveraged neural network techniques to analyze vehicle movements to improve machine operator behaviour.
Finally, A fully integrated and automated “intelligent mine” with the potential for being run entirely remotely gives us the opportunity to mine and extract minerals in areas of the world, and beyond, where humans simply cannot exist.

Related Articles from Nikolas Badminton:

How To Become A Centaur (Intelligence Augmentation)

Future Trends – Robots Are Destroying Jobs

Hire Nikolas Badminton, Futurist Speaker