Automation and Asteroids: How Rio Tinto is Changing the Future of Mining
When you can remotely operate an autonomous mine from pit-to-port, the sky is no longer the limit
With increasing urbanization and economic growth in developing markets, as well as the need to upgrade infrastructure in established markets, demand for minerals has never been greater. Unfortunately, most low-cost, easy-to-access (so-called ‘Tier 1’) mineral deposits have been depleted, so producing metals like iron ore has never been more difficult or dangerous.[1]
Physical challenges aside, the economics of mineral extraction are equally daunting. In commodity industries, the price of goods is set by the balance of supply and demand. Oscillating periods of over- or under- supply drive multi-year price cycles. Iron ore, for example, cost $180 USD/ton in 2011 when producers struggled to satiate Chinese demand for steel, but has since dropped to $50 USD/ton. In this environment, the key to capturing enduring value is to be the producer with the lowest marginal cost of production, so each ton of ore has the highest possible gross margin.
How can miners achieve this? The key is productivity, and the lifeblood of productivity in mining is technology. Technology minimizes the variable cost per unit of production (mainly labor and energy), and removes bottlenecks in production to maximize throughput (and therefore maximizes the denominator over which fixed costs are unitized).
Across the mining industry, Rio Tinto is leading the charge in innovating its operating model through automation, big data analytics and machine learning. Rio is the world’s second largest mining company, operating in 40 countries and six continents, and one of the largest companies on both the LSE and ASX stock exchanges. Their digital transformation began in the mid-late 2000s near the peak of a global commodities boom, when Rio’s constraint on production was its ability to hire qualified employees to operate in very remote locations. For example, it was an industry norm for truck drivers to earn over $250,000 per year, and mines were staffed by a largely fly-in-fly-out workforce that lived in purpose-built camps for shifts 1-2 weeks in length.
In response to these pressures, Rio announced its plan for the ‘Mine of the FutureTM’. In 2008, it began trials of autonomous vehicles.[2] In 2010, Rio commissioned its Operations Centre in Perth to centrally manage its Western Australian operations. Today, more than 400 operators track 3D visualizations of every piece of capital equipment covering 15 mins, 31 pits, four ports and 1,600 km of rail networks. Real-time big data analytics track productivity and safety of infrastructure that rivals the scale of many small countries. In many ways, Rio is a case study of what multinational companies can accomplish through an ‘Internet-of-Things’ network. By 2015, Rio had launched fully autonomous trucks in four of its iron ore mines, with plans to expand these machines to all open-cut mines.[3] In the past five years, Rio has also launched autonomous drills and applied neural network machine learning to optimize the flotation processes that extract refined minerals from ore.[4]
The benefits of autonomous and data-enabled operations are obvious. Most important is safety: robots don’t get fatigued, their attention span does not vary, and their loss from industrial accidents (like falling rocks or explosions) pales in comparison to the loss of human life. Other benefits include lower operating costs—there are now fewer truck drivers on $250,000 per year—and increased throughput from less down-time due to shift changes, and less inter-operator variability. Greater use of remote sensors further provides rich data that machine learning algorithms can mine for continuous improvement. Collectively, this has added billions in productivity improvements.[5]
While further innovation is underway, the journey toward automation is not without its pitfalls. For example, Rio has spent over $500m to deliver autonomous trains over the past four years, but failed to deliver the project on schedule due to ongoing difficulties (despite the promise of an extra two hours/day of productive time, a prize with an NPV of billions).[6] Furthermore, the company has faced significant resistance from unions, as jobs shift from highly-unionized frontline work to non-unionized knowledge and supervisory roles like analysts, programmers and robotics engineers.[7]
The end game for Rio Tinto, and the mining industry generally, is end-to-end automation from ‘pit-to-port’. Much in the same way that wars can now be fought with autonomous drones supervised from thousands of kilometers away, mining is moving towards 24/7 robotic operations with human supervisors located in the safety of developed cities.[8] This not only removes the threat of injury and death (present in mining since time immemorial), but creates the potential to extract minerals from the ocean floor or below the Siberian permafrost in the near-to-medium term. In the long-term, the digital innovation born from skills shortages in the Western Australian labor market enables mineral extraction from asteroids, removing the trade-off between economic development and environmental damage. (785 words).
[1] Economist.com, “The lore of ore”, http://www.economist.com/node/21564559, accessed November 2016
[2] Cecilia Jamasmie, “The Mine of the Future Might be a Thing of the Past,” Mining.com, January 2010, http://www.mining.com/the-mine-of-the-future-might-be-a-thing-of-the-past/, accessed November 2016
[3] Andrew Topf, “Two Aussie mines start moving all their iron ore with driverless trucks,” Mining.com, October 2015, http://www.mining.com/two-aussie-mines-start-moving-all-their-iron-ore-with-driverless-trucks/, accessed November 2016
[4] Rio Tinto, “The Mine of the Future”, 2014, http://www.riotinto.com/documents/Mine_of_The_Future_Brochure.pdf, accessed November 2016
[5] Tess Ingram, “The Pilbara’s productivity push: drilling down to the small stuff,” Australian Financial Review, September 2016, http://www.afr.com/business/mining/the-pilbaras-productivity-push-drilling-down-to-the-small-stuff-20160912-greo6p, accessed November 2016
[6] Dave Stringer, “Creating the World’s Biggest Robots Proves Challenging for Miners,” Bloomberg News, June 2016, http://www.bloomberg.com/news/articles/2016-06-17/world-s-biggest-robots-delay-dims-zeal-to-automate-mine-trains, accessed November 2016
[7] Andrew Duffy, “Unions vs Rio: The Automation Battle,” Australian Mining, March 2012, https://www.australianmining.com.au/features/unions-vs-rio-the-automation-battle/, accessed November 2016
[8] Kathryn Diss, “Driverless trucks move all iron ore at Rio Tinto’s Pilbara mines, in world first,” Australian Broadcasting Corporation, October 2015, http://www.abc.net.au/news/2015-10-18/rio-tinto-opens-worlds-first-automated-mine/6863814, accessed November 2016
This kind of progress is inspiring in the context of our insatiable need for extraction of minerals from the Earth. While operationally Rio Tinto is certainly leading the way towards automation, through investments that will pay back in groves to the company and save lives, one has to also scrutinise the operations beyond the mine itself. There are still substantial social costs to mining. This week two senior executives at RT got the boot for being complicit in bribery in Guinea. Furthermore, such automation is likely to also result in widespread unemployment in the local area. Should the company do more to mitigate these problems or is this a necessarily evil given the market for these minerals? Are there measures that regulation and innovation should focus on on the demand side?
Very interesting post, Nick. Mining is an extremely interesting area, in which new self-driving car/ truck technologies that are being developed globally can be employed to increase safety and drive productivity.
I think there are great benefits to using self-driving technology in open pit mining as you mention in the case of iron ore mining by Rio Tinto, but even greater benefits if technology advances sufficiently to be able to employ self driving trucks in deep underground mining. Whilst it is more difficult to control a truck underground with poor visibility and weaker GPS signals, Volvo has recently tested the first fully self driving truck in underground mining at Boliden’s Kristineberg Mine in Sweden. In the future these trucks can be equipped with sensor technology that would allow them to communicate geological, pressure and weather data to surrounding machinery and databases for analysis by scientists and mining supervisors.
These are very promising developments for the future of mining.
Related article:
https://news.volvogroup.com/2016/09/07/volvo-first-in-the-world-with-self-driving-truck-in-underground-mine/
Video:
https://www.youtube.com/watch?v=JwhyoUyJNoY
Great article Nick. I was aware that mining companies were using GPS-enabled systems to track performance but what Rio Tinto is doing with automation is very impressive, specially with the application of self-operating machines. I wonder how the core competencies of mining companies are going to change as they shift towards the “end-to-end automation from ‘pit-to-port’” you mention. Today much of the knowledge on mining operations lies with the on-site engineers and managers with a lot of experience. However, as the digitization transformation continues, it will more often be the case that the most relevant knowledge lies in engineers located the remote Operation Center with expertise in data analytics and robotics. With that in mind, do you think that mining companies will start to resemble technology companies with huge R&D focus?
Interesting article Nick – wholeheartedly agree that safety is the most important positive outcome from these advances in automation. I would be interested to see how industry safety statistics change (TRIR, fatalities, etc.) as more of this this kind of automated technology penetrates heavy industry. An interesting analog is the airline industry, which has become heavily automated. The report below found a statistically significant decrease in aviation accidents between 2000-2010 due to automation. If this trend is transferable (and I think it certainly is), worker safety will continue to improve in future years.
http://commons.erau.edu/cgi/viewcontent.cgi?article=1107&context=publication
Thanks Nick! Just wanted to add that this technology is (and I think that we can almost speak with certainty at this point) also going to be extremely important in the future situation where human labor is a very rare and precious resource – namely the early stages of colonizing the Red Planet and beyond.