In 2010, a 40-foot, 22-megawatt semi-autogenous grinding (SAG) mill set a world record for grinding power when the Esperanza copper mine put the
device, manufactured by FLSmidth, into operation. With low-grade reserves and a high daily throughput of around 90,000 tonnes, three-year-old Esperanza
is impressive, but may not the best model for future mines around the world.
“If you take a simplistic view of demand increase and declining head grades over the next 20 years, you might have to be processing four times the tonnage
that we currently process to make the amount of copper that the world needs,” says Joe Pease, CEO of Xstrata Technology. He notes that energy prices are
rising, energy security is falling, and available deposits are becoming more remote and more challenging to extract, further increasing the energy needed
for each of those tonnes.
If the only problem were accommodating higher tonnages, more grinding power in a standard SAG mill-ball mill processing circuit would suffice. And that is
part of the answer. Harri Lehto, technology manager, grinding processes at Outotec, says giant tumbling mills are “more or less a must” for any equipment
manufacturer wishing to be taken seriously. Incremental increases in mill size have brought the maximum rating up to 22 megawatts, while 42-foot,
28-megawatt SAG mills have been designed (though not installed).
But can they go bigger? “We have been stuck at 40 feet for over 10 years,” says Steve Morrell, a long-time comminution researcher and current managing
director of SMC Testing Pty Ltd. The bigger SAG mills use gearless drives, and a number of recent drive failures have led to a general reluctance to push
the envelope beyond the 28-megawatt maximum.
The looming problem is not tonnage in itself but in the cost of tonnage, exacerbated by tumbling mills’ inherent inefficiency. Rotating containers tossing
rock and steel spend most of the power they consume on generating heat. Ball mills use as little as one per cent of their energy draw to break rock.
The industry is discussing – and in some cases adopting – more complex but potentially more effective and energy-efficient milling circuits that do not
rely exclusively on scaling up tumbling mills. In terms of energy use, it can be more cost-effective to prepare ore for the ball mill by sending it through
multiple crushing stages rather than recirculating it through a SAG mill. The same staged approach to grinding divides the circuit into coarser- and
finer-grinding equipment to more efficiently recover the materials available at different sizes.
The last two decades of mill development have produced more equipment suited for size-specific crushing and grinding. High-pressure grinding rolls (HPGR),
typically used in the final stage of crushing, compress the coarse ore between two cylinders. They may use 10 to 20 per cent less power than SAG mills.
On the other end of the size spectrum, stirred mills grind fine fractions using a rotating shaft within a stationary shell. Stirred mills cut energy use
because only the shaft and discs need to be rotated, and because the mills can use small, high-velocity grinding media with a much greater surface area for
grinding. They are typically used to regrind ore once it has passed through a ball mill.
“These technologies have been discussed for many years,” says Walter Valery, global senior vice-president at Metso Process Technology and Innovation. “For
example, we proposed a circuit flowsheet with HPGR followed by stirred mills in conjunction with high-intensity blasting as an energy-efficient alternative
over 10 years ago. However, only recently are we seeing HPGR technology considered in most prefeasibility and feasibility studies.”
Costs and benefits
Both HPGR and stirred mills have made inroads. One of the two most widespread suppliers of stirred mills, MetsoMining and Construction Technology, has a
total of 39 installations in Canada. Paul Cousin, vice-president of metallurgy at Agnico Eagle Mines, says his company is considering using stirred mill
technology to regrind the ore from its LaRonde mine in Quebec. What he has heard suggests that the mills yield a better end-product, adding not just energy
efficiency but overall cost payback.
Nonetheless, the inherent conservatism of the industry slows adoption of proven technologies, says Jonathan Allen, product manager for stirred mills at
Metso. “Everybody in mineral processing knows ball mills,” he says. “So when a new supplier brings in a new technology, no matter what it is, if it’s
unique to one supplier, you’ve got everybody else out there saying it’s a bad idea. The market penetration just takes a bit of time in our industry.”
Capacity is another concern, says Allen. Stirred mills were first used for regrinding so they have a small throughput relative to ball mills, but as their
capacity has increased with the move to whole ore grinding, so has interest. In 2008, replacing one 12-megawatt ball mill would have required a series of
anywhere from two to eight smaller stirred mills. Five years later, Metso is working on a 4.5-megawatt version of its Vertimill, or vertical stirred mill,
and Xstrata Technology offers an eight-megawatt horizontal IsaMill. Stirred mill installations have multiplied in the last few years. This includes
large-scale adoption by Anglo Platinum, which extensively uses horizontal stirred mills for tertiary grinding as well as for regrinding.
While HPGR technology at first promises large increases in energy efficiency, more holistic evaluations of HPGR erode the cost savings that “energy
efficiency” measures imply. They may have lower operating costs and offer certain savings – for example, no need for steel grinding media – but HPGR also
require additional auxiliary equipment, precrushing, extra screens, conveyors, storage and dust extraction. Brian Putland, president of Toronto-based Orway
Mineral Consultants, explains that those extra items can add up to a high-capital investment compared to SAG installations. Capital costs reflect energy
costs as well, argues Putland: A full evaluation of the energy used would include the manufacturing and shipping of equipment and wear items.
Alan Muir, vice-president of metallurgy at AngloGold Ashanti, says the company considers HPGR at any new project but has only installed one such circuit,
at its new Tropicana mine in Australia. “That decision was really driven by the very hard nature of the ore and the cost of on-site power generation, which
is extremely high,” he says. Tropicana runs up power costs of 27 to 30 US cents per kilowatt-hour.
“If we take South America, where there’s a lot of hydropower production, there the cost is typically in the region of 9 to 12 US cents per kilowatt-hour,”
adds Muir. “So there, it doesn’t really stack up. You would spend more money on capital equipment and not have the savings on power that you need to offset
In Canada, cheap hydropower puts energy costs lower on the priority list. About 90 percent of Putland’s clients go with a conventional SAG and ball mill
set-up after considering alternative methods. Even Far North projects, where power is expensive, find that the costs of covering, heating, operating, or
maintaining additional equipment can outweigh the energy benefits.
Cousin says HPGR were among the options considered but discarded at Agnico Eagle’s evaluation-stage Meliadine gold project in Nunavut. “It was our belief
that HPGR could be well-suited on an energy basis, because of the high cost of producing power out there,” he explains. “But to our surprise, the design of
the overall comminution circuit including an HPGR portion proved to be not as efficient, in terms of overall economics, as a more conventional approach of
A better circuit
Steve Walters, research director of the industry-funded project Cooperative Research Centre for Optimising Resource Extraction (CRC ORE), stresses that
focusing on the energy efficiency of equipment misses the point. “It’s like rating a washing machine and not checking that it cleans the clothes,” he says.
“We can argue about the efficiencies of HPGR comminution processes versus a SAG. That’s not the real question here. The performance metric shouldn’t be the
efficiency of the activity; it should be the useful output.”
By “output,” Walters does not mean throughput: he means metal, the ultimate unit of success. At the University of Queensland in Australia, CRC ORE and the
Julius Kruttschnitt Mineral Research Centre (JKMRC), researchers are working on approaches that process less rock and recover more metal: in other words,
how to do as little work as possible on as little ore as possible.
Malcolm Powell, chair of the Anglo-American Centre for Sustainable Comminution at JKMRC, suggests that mines stop blending different ore grades and start
separating different grades into different streams, to be processed in different circuits. The tools to do this already exist; bulk grade detectors can
already sort out ore, and some large operations already have multiple ore streams, simply because they have too much
production volume to process in one mill. That presents an opportunity to apply what Powell calls “flexible circuits” that respond to a specific ore, via
conventional or novel milling methods as appropriate.
AngloGold Ashanti is following this approach at its projects in development. “Just blending everything together through a single circuit is probably not
optimal,” agrees Muir. “The trick will be to develop circuits which have the flexibility to adapt to different ore types and maintain optimal processing
even when the ore changes.”
Ore body knowledge is a critical part of this approach, and that is one reason that it has not been more readily taken up, explains Pease. He estimates
that knowing what can be done with the ore is about 80 per cent of the work needed. With that in place, each circuit ought to require less equipment
because it has been designed with clinical precision. But the individual nature of the test work and resulting flow sheet can be off-putting.
“In theory, it’s there,” says Pease. “We have the drill core. We have quantitative mineralogy. We have diagnostic crushing and comminution and laboratory
tests. We can really map the ore body and map the metallurgical response and custom-design a flow sheet for it. I’m not sure we always make use of that as
much as we can, because it seems expensive and time-consuming and takes a fair bit of expertise. It doesn’t suit fast-track engineering and the flow sheet
that comes out of that is individual. Perhaps from the position of the board of directors, they’re saying ‘Unique design sounds like risky.’ You need to be
able to explain to them that it means custom-designed to be lowest cost for this ore, with savings far outweighing the investment in time and design.”
Integrating the mine site
What any mine can and should do, many agree, is to improve cross-silo communication. That is critical to Pease and Walters’ suggested strategies, which
focus on doing work as early as possible: blasting selectively to produce a better feed, using pre-concentration and crushing to reduce the work done by
more energy-intensive grinding equipment, and producing the highest grade possible in concentrate, since smelting uses far more energy to remove the same
impurities. At Agnico Eagle, talk within the operations team now includes discussion of how drilling patterns could increase costs on the mining end but
reduce the cost of operating comminution equipment.
Powell says the potential benefits of adopting more complex controls are apparent and well-supported on a simulation level; the challenge is to demonstrate
and quantify those benefits in practice. Given the risk-averse climate, only a few mines have volunteered to try out JKMRC initiatives. But Powell believes
that the industry as a whole is moving toward a more integrated, mine-to-smelter perspective. “It’s not obvious in the way businesses are run yet,” he
says, “but the way we’re talking to industry and the way we think about the problem now is much more a systems approach.”
The upside to the downturn
In the midst of the last mining boom, when skills were scarce and services expensive, redesigning comminution circuits took a backseat to getting projects
up and running. Mining companies’ current financial challenges have put a new premium on operating what they have with greater efficiency.
Grant Ballantyne, a research fellow at JKMRC, has seen attitudes change within 18 months. During a workshop held by the Coalition for Eco-Efficient
Comminution in 2012, when most metal prices were rising or had hit a plateau, attendees emphasized throughput as the financial driver of their comminution
choices. At another CEEC workshop in July 2013, all 16 attendees – from vendors, operations and engineering companies – came seeking efficiency and
productivity increases. “The push for operational efficiency seems to be increasing,” he says.
But Walters is not positive that the downturn will inspire innovation. “The mining industry is right on the cusp of change,” he says. “Some mines have
started to take a bit more of a flexible approach to how they deal with their ores, but there are very few of them. So given the change in the industry
dynamics, there’s a big issue here as to whether they will retreat to what they think is safe, or whether they’ll evolve into something which is going to
be more efficient, but also more profitable.”
While major innovations are often left to major companies that can afford to open themselves up to such risk, Valery’s experience suggests that solutions
could come out of smaller mines as well. “Smaller companies often have more difficulties raising capital than larger companies and therefore need to lower
capital expenditures and operating expenditures to get their projects off the ground,” he explains. “In order to do this, they are more receptive to
innovative or alternative solutions, which are riskier than the conventional ones.”
The future of comminution
What will a standard circuit look like in 10 years’ time? “The standard will be that there is no standard,” Pease answers. “I think that’s sort of the
problem at the moment, is that for lack of resources, we design a standard plant. The standard should be that the plant is custom-designed for exactly what
this ore body needs. And so the standard will consider first of all how much fragmentation is done in the mine; it will consider what can be done with
pre-concentration to remove really coarse gangue early on; it will then consider a stage grinding and flotation flow sheet to minimize grinding energy.
It’ll use the least amount of grinding energy on the lowest possible tonnage. And that standard approach will build, I believe, smaller, more efficient
Can comminution be done away with altogether? Muir thinks that is a question worth posing, and he plans to address it in a keynote lecture next year. “It’s
a little bit provocative to say, however, we are being pushed into the corner by rising costs, dropping feed grades and higher throughput rates, and all
we’re trying to do is tweak the existing technologies that we have, where we should be really spending more money on exploring new technology based on
different science,” he comments. HPGR use compression force to break down particles rather than the typical impact forces employed in tumbling mills. “If
compression is more efficient than impact (at finer sizes) we need to explore what is more efficient than compression. We have also started to look at
whether we can do in situ leaching, which would eliminate the need for mining and surface operations, grinding and all that kind of stuff. It’s not a
technology I can switch on tomorrow.” But, he says, “I think it’s time we started looking at that as an opportunity worthy of serious consideration.”