From autonomous mining trucks to big data analytics — how long will it take until mining is a fully automatic process carried out from a bunker in Perth?

Mining has a reputation of being a low-tech industry, but in recent years has harnessed the full gamut of technology to drive productivity gains far beyond that achievable by humans.

While the basics remain in play and the wisdom of experience is still valued, it’s those with protractors rather than hard-hats who are dictating how mining operations are run.

It’s generally because mining operations are larger than they’ve ever been before, so the scope for cost savings or overruns are therefore magnified.

If you can achieve a cost saving of 1 percent on a $500 million mining operation from a $2 million investment, it’s worth it.

If you can achieve a cost saving of 1 percent on a $5 million mining operation, a $2 million investment simply isn’t worth it.

But as the big end of the mining space use more tech, the cost curve will come down.

As technology has evolved and the business case for investment in new technology to achieve efficiencies has strengthened, the average mining operation now has more tech than you can poke a stick at.

Here are three of them.

1. Autonomous Mining Trucks

Early last year, Rio Tinto (ASX:RIO) hit a milestone which was enough to trigger somewhat of a nihilistic crisis.

Its autonomous mining trucks moved their one billionth tonne of material in the Pilbara.

Scary stuff, if you’re a miner — but a game-changer if your a mine proponent.

The autonomous mining trucks operated 700 hours longer than their human-driven counterparts and had a total unit cost 15 percent lower.

At the start of last year it was reported that it had 80 such trucks in the field, but was moving to increase this to 140 by the end of this year.

Obviously autonomous mining trucks are only really the domain of huge miners at the moment as the capex required to outfit a fleet can be expensive.

But new technology always requires an initial user base to take a punt in order to drive down the cost-curve and work out a few of the kinks in the field.

So one has to wonder how long it will take for all mining trucks to be autonomous.

After all, they operate in some of the harshest areas of the planet and are a major cause of incidents for miners.

2. Machine learning

Take Woodside Energy (ASX:WPO).

In recent years it’s made massive investments in technology usually reserved for those working in Silicon Valley.

At one point, it was running the single largest commercial instance of IBM’s Watson machine learning tech in Australia.

Watson is essentially able to aggregate huge amounts of data and then make a prediction based on that data — and the more data you feed it, the more accurate it becomes.

How does this play out?

Take the average LNG plant for instance — with its incomprehensible systems and intertwining tubes. They’re generally large, incomprehensible projects — and no one person knows how to operate the entire plant.

Instead, knowledge is spread out among separate people — many of whom are senior management and are expected to transition out of the business in the coming years.

So what hope do new people coming in have of interpreting this vast system?

They don’t.

It’s much like making a cake when you only have one part of the recipe.

Instead, you could feed a machine learning system such as Watson all the technical information and reports so it becomes the repository of learning in the organisation.

It can then respond to queries from newer engineers about what they’re looking at, and how to fix a certain problem.

Previously, this would involve finding the wiser heads in the organisation — if they’re even there.

Now, the collective intelligence of an organisation sits with the machine.

What makes it different from a database of technical documents, however, is the ability for the AI to spit out insights based on natural language processing — meaning engineers can literally ask a question the same way they would have asked it of a senior engineer.

Of course, the secret sauce of what makes a machine learning solution such as Watson effective is data…

3. IOT and predictive analytics

One of the huge pieces of work done in the mining space over the past decade is in attaching sensors to pretty much everything — and it’s done for a reason.

Previously, engineers on huge mining operations would need to go out and take sensor readings on all sorts of crucial equipment, measuring things like how much pressure is being built up in a valve or corrosion on a pipe.

They knew from previous experience that should the pressure be built up too much, then it could lead to catastrophic equipment failure somewhere down the line.

It was hard to tell exactly when something would go bust — but by adding IOT sensors to equipment to feed back data on a whole range of pipes and valves, the operator can look at every variable before, during, and after a disaster and put measures in place to prevent it happening again.

Looking to Woodside again, its principle data scientist Elsa Jordan told a conference in 2016 how this might play out.

A production facility strips out acid gas and carbon from gas before its liquefied at one of its LNG plants — but this process can lead to production issues. What’s more, these issues can be hard to detect before it shuts down production, costing millions of dollars.

But thanks to IOT sensors it did have a bunch of pressure data it could funnel into a predictive analytics engine.

Was there a change in pressure around the stop-production event? Does a correlation imply causation? Can the way pressure is changing at any moment effectively predict a looming stop-production event?

“The question was whether we had the pre-event data that could warn us in the future, and the answer was yes,” Jordan told the conference. “But it’s encoded into millions of rows, from readings across thousands of sensors, to logs and regular data.”

Woodside was able to apply a predictive algorithm to the data, and come up with a probabilistic model of when the problem would occur based on the real-time data generated by the IOT sensors.



The mining business is big business. The sheer scale of a mining operation is huge – and as such looking for ways to innovate to drive efficiency gains is also big business.

So, tech like autonomous mining trucks and AI could provide the next big wave of gains for the sector as a whole — and we’re just starting to see it play out.