By Bryony Collins, Bloomberg New Energy Finance editorial team. This article first appeared on the Bloomberg Terminal and is available to BNEF clients on the web.
Anglo American Plc is using digital technology and machine automation to improve the productivity and safety of its mining operations. Companies that have digitized their technical equipment have “generally seen around a 30 percent improvement in their business – made up of about 15 percent in productivity and 15 percent in cost savings”, Tony O’Neill, technical director at the global mining company, told BNEF.
The U.K.-based miner of commodities such as nickel, coal and precious metals is using “digital twins” to optimize its mining fleet, including applications to track the performance of haulage at its Los Bronces mining site in Chile and at a 500 kilometer pipeline in Brazil, O’Neill said in an interview at Anglo American’s London headquarters on February 5.
Digital twins are virtual models of a technical process that allow companies to analyze data and improve the efficiency of how a piece of equipment works in the real world. Anglo American implemented its first of these virtual models 18 months ago, and plans to install them across its entire business “pretty quickly”, said O’Neill.
The miner is also developing robotics that have the potential to remove people from underground mining altogether, bringing significant safety benefits and productivity gains. Rock cutting technology and robots that can “chisel rock at a rate equivalent to a human” are being trialled using machine learning. This change could increase productivity by as much as 30 percent, said O’Neill. [Productivity is defined as tons produced per full-time employee].
Due to declining quality of ore grades, energy consumption in mining has increased by 16 times since 1900, as more energy and water is used to produce the required amount of metal. Machine learning can help increase the efficiency of these operations by improving the precision of mining, so that less waste, water and energy get used, according to O’Neill.
“Innovation in mining has traditionally meant scaling up – bigger trucks, bigger shovels – to increase capacity”, he added. However, “with robotics, the smaller the better: lightweight and modular, robots can perform multiple tasks in any terrain.”
O’Neill discusses the challenges and opportunities in digitizing large mining operations in the following interview.
Q: How has the mining industry changed in recent years?
A: The mining industry has expanded significantly [with larger mine sites and heavier trucks]. But the amount of energy being put into this expansion is unsustainable. For example, in 1900, to produce 40 kilograms of copper, 2 tons of rock were mined. Today, because of declining grades, to produce the same amount of copper, material movement and energy consumption has risen 16 times and water consumption has doubled (per unit). Clearly, this trajectory is unsustainable.
While efficiency is implied, I am convinced it is about precision – precision to target only the metal or mineral, with radically less waste, less water and energy inputs and less footprint.
Q: How is Anglo American improving the sustainability of its mining operations?
A: The industry has the opportunity to change its footprint, [in terms of reducing its energy, water and land usage requirements]. At some point in the future, there will be mines that basically don’t have a footprint.
The day is coming when mines won’t use water. We’re investigating the use of chemical replacements for water. Water sent to tailings ponds [dams where waterborne refuse from mining is pumped] often represents the largest water loss at a mine. In the future, tailings ponds will disappear and be replaced with dry tailings stacks [dry deposits].
In terms of power, we’re getting to a flip point on renewable sources. The size of the facilities that we require are really quite large [and] some countries won’t allow us to go off-grid. So we’re investigating our pathway forward and clearly renewables for us are of great interest.
[The BNEF report Miners and Renewables: Will They Dig It? looks at the potential for renewable energy to power mining sites. Clients can access it on the Terminal or on web.]
Q: What is the average power demand of a mine site?
A: It would be 100 to 150 megawatts — processing and grinding rock is very energy-intensive.
Q: That amount of energy could easily be supplied by a solar farm..?
A: [It’s a matter of finding the energy storage] that we require to be able to operate around the clock. But that’s all changing, and we certainly have an interest in hydrogen power, especially through our production of platinum and palladium. Hydrogen cells give higher range than that currently provided by battery technologies — you typically get double the range [energy density] out of a hydrogen cell than you do a battery.
Q: What work have you been doing with digitizing your mines?
A: We’re already starting to use digital twins to optimize our mining fleet. [We use them to reduce diesel consumption on our trucks] and we are also putting digital twins across pipelines, smelters and refineries… targeting to put right across our business.
Today, we use analytics for real-time drilling analysis, hyper-spectral core imaging, and geological modeling software using 3D and virtual-reality technology to generate and interpret predictive data models.
Our latest work uses customized learning algorithms to predict control parameters required by our plants. Another important application is condition monitoring and predictive maintenance.
The logical end-point is a fully-integrated, systematized and self-learning operation that will help remove the uncertainty and huge variability that characterizes mining today.
Where people have gone digital, they’ve generally seen around a 30 percent improvement in their business – made up of about 15 percent in productivity and 15 percent in cost savings. That’s what we’re targeting.
Q: How would you implement a digital twin to make a mining process more efficient?
A: A machine, whether it’s a pipeline, a truck or a smelter, has a theoretical way of behaving and as soon as you can start to measure the behavior as variants, then we can intervene.
We will use our digital twin to learn from the physical machine and then to actually change its behavior [from simulating how it acts in a virtual sense]. By using a digital twin, we can make our processes more effective and efficient.
Q: Give some examples of where you’ve implemented digital twins?
A: We’ve used digital twins in Chile at the Los Bronces mining site, where we implemented them into the haulage fleet to track the actual performance. We’ve got a 500-kilometer pipeline in Brazil and we’ve put a digital twin onto that. So all the key pieces of our business we’re starting to tweak. We’ve established a data analytics group in the company to lift our level of involvement and expertise in this space.
Another area that is quite interesting is geology, where through imaging we are now able to understand the grade and behavior of some ore bodies.
Q: What steps are you taking to reduce CO2?
A: Just by changing the cleanliness of fuel, we reduce the energy consumption of our trucks by about 8 percent. There are other trends such as steepening the walls of mines, which flow through to energy and footprint. It’s about looking at novel ways to steepen the walls of mines [so that less energy is used for excavation]. That alone could knock 15 to 20 percent off the industry’s energy consumption. These are all angles that we’re working on.
Q: How are you using robotic, autonomous technology to remove the need for blasting in mining?
A: We are working with small robots to get them to chisel rock at a rate equivalent to a human. Innovation in mining has traditionally meant scaling up – bigger trucks, bigger shovels – to increase capacity. With robotics, the smaller the better: lightweight and modular, robots can perform multiple tasks in any terrain.
At our underground mines in South Africa, we’re developing rock cutting to allow us to break rock in a continuous fashion without putting people at risk. They will also be run with machine learning.
Those trials have been running now for two years and are looking promising. The safety, efficiency and productivity gains from a continuous mining process in hard rock are potentially significant and could put underground mining on a more even footing with open pits.
It’s about developing mining technology that can actually cut really hard rock without destroying itself. This differs from a typical mining process where people go underground, drill holes and put explosives in them, and then after 8 hours they’ll blast. And you can’t re-enter for a period due to smoke and other debris. By using cutting technology, it’s possible to continuously cut without starting and stopping, or having the re-entering problem.
Q: By how much do you think this could improve productivity [tons produced per employee]?
A: We’re looking for 20-30 percent increases – step-change stuff.
Read this BNEF report about how Barrick Gold Corp. is embracing digital technology to make its mines safer and more profitable.