A technological revolution is underway in the segment of the mining industry that is closely linked to metallurgy — working methods that have remained unchanged for decades and centuries are giving way to new ones. Before our very eyes, robots are becoming human assistants and polluting industries are being transformed into clean ones. A number of trends are changing the face of the world’s most important industry before our very eyes.
The brain is good, but so is artificial intelligence.
The exploration and evaluation of mineral deposits has traditionally been carried out by humans, and it seemed that human intelligence would never be replaced by artificial intelligence. However, more and more geologists are using AI to study mineral deposits; they do not see AI as a competitor, but rather as a reliable assistant that can simplify and speed up the interpretation of various data (geophysical, mineralogical, geochemical, etc.) and make predictions based on it.
Although the use of artificial intelligence in geological prospecting is generally still in its infancy (although the first scientific papers on the subject appeared as early as the late 1980s — early 1990s), there are already a range of information processing methods — expert systems, machine learning, and others. The most important of these is machine learning, i.e. the use of self-learning algorithms that can analyse large volumes of geological data involving multiple sources and types of information. They can identify deposits that have escaped human scrutiny or are hidden by thick rock formations.
The possibilities offered by artificial intelligence could not fail to attract the attention of the world’s leading mining companies. In 2021, for example, BHP Group and KoBold Metals agreed to jointly use artificial intelligence to search for copper, nickel and cobalt deposits. The partners have decided to use it in Australia, and it is likely that artificial intelligence will be used on BHP Group projects around the world in the future.
In 2022, Canada’s MacDonald Mines Exploration used machine learning to find an interesting copper-gold-silver mineralisation near the shore of Lake Guron. The identification of a large mineable deposit cannot be ruled out in the future.
Headland Mines
The prospects for AI in mining and mineral processing are equally good, paving the way (in combination with the Internet of Things) for the creation of ‘smart’ mines. Much of the AI will be needed to keep process equipment up and running, providing operators with real-time information on the status of equipment and machinery to identify failures and malfunctions: sensors collect data on operating parameters and transmit it to a central system. AI analyses the data and provides operators with information that indicates when something is wrong. Cloud technology, on the other hand, ensures the security of the data, while blockchain speeds up the transfer of data.
Artificial intelligence is also suitable for ore refining management. In 2019, Norilsk Nickel used it at its Talnakh concentrator to quickly track rock debris that could lead to equipment damage and cut the conveyor belt if the debris got into the crushing drum.
A year earlier, Rio Tinto approved the construction of the 43 million tonne per annum Koodaideri iron ore project in Australia. The $2.6 billion investment has become a pioneer in Rio Tinto’s smart mining operations, using autonomous drilling rigs, trucks and data analytics to optimise production and reduce downtime. Last year, Koodaideri, now renamed Gudai-Darri, began delivering raw materials and is expected to reach full capacity by 2023.
The smart mines will be well supported by drones that track the terrain in their area. US-based Freeport-McMoRan uses them to monitor the slope angles of quarry walls to prevent collapses. The information gathered by the drones is used to manage the removal of overburden and reduce the amount of waste rock mined.
Where it’s dark and damp
Can you imagine a mine or quarry without people in it when it is in operation? It may seem impossible. But this view is deeply flawed. The possibilities for using robots in mining are vast, and their use is driven by the high demands on human labour and the need to ensure human safety: mines can accumulate methane and explode, the air is often damp, and there are occasional cave-ins. It is therefore better to replace people (where possible) with unassuming machines.
One option is to use robots to assess the condition of underground mines. In 2014, the Freiberg Mining Academy launched the Mining-RoX project, which aims to use robots to map and monitor the mining environment. Two machines, Alexander and Julius, named after famous German scientists, were created.
They were tested at the Reiche Zeche mine, which had previously been used to mine silver ore. Each vehicle had a different role. Julius acted as an assistant surveyor, transporting heavy equipment and collecting the necessary data with portable measuring instruments. Alexander took care of the 3D mapping of the mine tunnels using a range of cameras and lidar sensors.
Trials with both robots produced good results, but also revealed a number of challenges for their use underground. These included a lack of GPS capability (as satellite signals cannot pass through opaque obstacles), rough terrain (which could block or even damage the vehicle) and the need to protect it from corrosive mine water.
However, their use makes it possible to automatically create accurate and economical 3D models of mines, greatly simplifying the planning and execution of mining operations. It is no coincidence that Norilsk Nickel began testing the Mark robot, developed at the Moscow Institute of Steel and Alloys, in September 2022.
Mark is characterised by high mobility, thanks to the design of its wheels: they all have their own motors, vertical rotation axes for turning and systems for changing the ground clearance from 20 to 45 cm. The robot can travel at speeds of up to 4 km/h and is designed for surveying and monitoring the air in the mine.
On a motorway without a driver.
As well as monitoring mines and quarries, the robots can also be used to transport ore. Machines don’t need clean air, good visibility or a lunch break. Although human drivers can do a good job, autonomous trucks working 24 hours a day can drive the exact route every time without getting bored, tired, resting, or making mistakes.
Japan’s Komatsu was a pioneer, testing five autonomous trucks at the Radomir Tomic mine in 2005 and beginning to build a fleet of them at the Mistral mine in Cabriela, both owned by Chile’s Codelco, in 2007.
Since then, the world’s biggest mining companies, including Rio Tinto, Vale and Fortescue Metals Group, have shown interest in operating them. There have been a few hiccups, most notably in 2008 when the machines were found to be malfunctioning at Cabriela Mistral and Codelco asked Komatsu to repair them.
Experience has shown that unmanned vehicles can increase productivity by 15 to 20 per cent, while reducing tyre wear by 5 to 15 per cent and fuel consumption by 10 to 15 per cent.
However, there have been concerns about human safety, which is why dump trucks supervised by operators in offices are also being developed. Norilsk Nickel, for example, has launched a project to use unmanned dump trucks at its Skalisty mine in 2021. This is currently unprecedented in Russia.
The prerequisites for its implementation were as follows: every mine has processes, such as blasting and ventilation, where personnel cannot work in the same place as the process. As a result, time is lost that could be used to transport the blasted ore to the concentrator. Unmanned dump trucks, on the other hand, can work at exactly the same time as any underground workings are being blasted and ventilated.
At Skalisty, an approximately 3-kilometre-long haulage tunnel at a depth of 850 metres was chosen as the break-in area for the machines. Switches, Wi-Fi access points and other network equipment were installed there, and the trucks were fitted with control units, laser scanners, video cameras and Wi-Fi antennas, as well as various sensors and detectors.
The project included training staff, identifying bottlenecks and key factors affecting the efficiency of ore delivery, and reducing the number of miners.
The project has shown that the unmanned trucks work 1.5 hours longer per day than their human-driven counterparts. They are loaded and unloaded remotely. As a result, the amount of material transported increases by at least 1%, or 21,000 tonnes per year.
The people on Skalisty are left to do the intellectual work, armed with modern gadgets. On Skalisty, they use lasers to measure, download the data onto tablets and use them to create three-dimensional models of the mine shafts. This makes the job a lot easier and increases the accuracy of the measurements.
…and on the rails.
There’s also a place for robots on the railways that carry raw materials from mines to ports. Between 2012 and 2019, Rio Tinto implemented a massive Autohaul project in Australia, spending almost $1 billion to create an autonomous rail network that transports about 1 million tonnes of iron ore every day.
The network consists of hundreds of heavy trains. Each train is 2.4 kilometres long and consists of 2 to 3 locomotives and 240 cars, carrying a total of 28,000 tonnes of raw materials. The trains run between Rio Tinto’s 17 quarries and the ports of Dampier and Cape Lambert. The average turnaround time is 40 hours. The wagons are loaded and unloaded automatically, and the train runs autonomously. The conductor only operates the train as it passes through the port area.
No gas emissions
Transporting ore by dump truck or train involves high fuel and energy costs. Emissions into the atmosphere are also significant: an estimated 28,000 dump trucks currently in use worldwide produce up to 68 million tonnes of carbon dioxide per year, not including the toxic gases produced by the combustion of diesel fuel.
To reduce their negative environmental impact and energy costs, several alternative non-petroleum fuels have been proposed — liquefied and compressed natural gas, ethanol and methanol, and hydrogen.
Hydrogen is the most environmentally friendly of these, producing environmentally friendly water as a by-product. Hydrogen can be produced by either steam conversion, methane pyrolysis, or water electrolysis.
The use of hydrogen in mining trucks has been a focus for the UK’s Anglo American, which has agreed with Engie to build the world’s largest hydrogen-powered mining truck. The prototype was unveiled in 2021 at the Mogalakwena mine in South Africa, which mines platinoids.
Tests will determine whether the truck will replace Mogalakwena’s fleet of 40 diesel trucks and is expected to reduce emissions by 80%, saving the company up to 1 million litres of diesel fuel a year.
Sun and wind
The use of hydrogen reflects a trend in recent years towards the use of renewable energy sources in the mining industry.
In 2019, Rio Tinto announced its intention to reduce its carbon footprint at the Kennecott Utah mine in the US by 65% by purchasing 1.5 million MWh of green electricity from Rocky Mountain Power, which operates wind farms.
This was followed by South Africa’s Gold Fields, which signed an agreement with EDL Group to supply its Agnew gold mine in Australia from a hybrid power plant powered by solar, wind, and diesel. The plant has a total capacity of 56 MW, with 18 MW coming from wind and 4 MW from solar.
In addition, Barrick Gold, OZ Minerals, BHP Billiton, Sandfire Resources, B2Gold, Antofagasta and other players have shown strong interest in renewable energy. According to Fitch Solutions, around 1GW of wind and solar capacity has been built at mining sites around the world in 2020, with a similar number in the planning or construction stage.
When analysing project information, it is easy to see that much of this is concentrated in Australia and Latin America, which have a high number of sunny days per year and strong coastal winds. However, both solar and wind power have drawbacks: the efficiency of the former depends directly on the angle of the sun, the amount of cloud cover and the time of day; the latter depends on the strength of the wind and the time of year. Solar power plants do not work at night, and wind power plants do not work when it is calm.
Curiously, the mining industry almost never uses geothermal energy. The only exception is a project by Australia’s Vulcan Energy Resources to extract lithium from heated groundwater in Germany and use the heat as an energy source.
In the future, the share of renewable energy in the mining industry’s energy mix will increase, but not everywhere. The main contributors will be North and South America, Australia and, to a lesser extent, Africa, Europe and Asia, due to different geographical conditions. 
