Article originally published in International Mining (IM), March 2019. Reproduced with kind permission.
In their time of need, mining companies are looking to pre-concentration and ore sorting solutions to upgrade their product feed. Dan Gleeson examines just what the industry has to offer
The world of ore sorting has become a lot more interesting in the past decade as mining companies have started to re-examine old concepts to stem the sector-wide grade decline and improve profitability.
Photometric, electromagnetic, radiometric and X-ray sensor technologies are, in some instances, enabling miners to cut the amount of waste material going into processing plants, or, conversely, boost the amount of ore fed into the comminution circuit.
When positioned correctly in a process flowsheet, a pre-concentration stage or sensor-based ore sorting mechanism can reduce material transport costs, meaning less waste is conveyed or hauled to the next part of the mining process.
All of this can, potentially, lead to reduced energy consumption, a smaller environmental footprint and increased profitability; goals the industry has been looking to achieve.
Mineralogy & heterogeneity
“It’s mostly driven by the mineralogy,” Cronimet Mining Processing SA Operations Director, Ruan Kroukamp, said in reference to how wide the applications were for sensor-based ore sorting in mining. While Kroukamp’s experience is mostly related to the application of X-ray transmission (XRT) technology, the statement rings true for most sensor-based ore sorting solutions.
TOMRA Sorting Solutions’ Business Development Manager, Christopher Robben, added to this: “The value (of ore sorting) is more dependent on the mineralisation of the deposit and liberation characteristics of the ore in combination with the value chain of the operation.”
Robben raises an important point here – the placement of the ore sorter in a mining operation has a massive impact on just how successful it is.
Place the ore sorter in the wrong place and the separation efficiency is nowhere near as high as it could be.
Metso explained: “Most mining deposits are naturally heterogeneous and lend themselves well to bulk ore sorting, but it should be implemented as early in the process as possible before excessive mixing occurs. Material presented to the sorter needs to have sufficient grade variability occurring in large enough batches of material for effective separation.
“But every time the ore is re-handled, transferred, crushed or blended, the degree of mixing increases; reducing the variability and thus the potential for effective separation of batches of barren gangue from ore.”
In terms of the bulk ore sorting process (more on this later in the Bulking up section), either in-pit or plant feed conveyors provide the best opportunities for sorting, according to Metso, adding that this enables the sorter to exploit the natural heterogeneity of the deposit.
And, the pre-concentration stage can be as varied as the sensors used to separate ore and waste.
Malcolm Guthrie, Lead Process Engineer, explained: “It is a broad area of processing that to some degree is an extension of grade control and incorporates any technique that removes waste based on identifiable/detectable characteristics, including particle size, colour, detectable minerals or elements and atomic density.”
Coal operations use a mixture of dense media separation (DMS) and magnetic sensors in their operations to remove the tramp metal, increase the value of their product and protect the screens, crushers and conveyors that process the material later in the flowsheet.
Radiometric sensors have been used in uranium operations to upgrade concentrate before it is processed into yellowcake, while research has shown photometric sensors could have their uses in the gold mining industry, in addition to other commodities.
And, of course, XRT sensor solutions have gone from being mainly used in diamond and industrial metal operations to potentially finding new homes in uranium, iron ore, coal, copper, gold, silver and phosphate mines, to name a few.
The ore sorting solution gaining most traction of late is XRT.
Most of the XRT solutions on the market recognise and separate materials based on their specific atomic density, allowing a high level of sorting purity irrespective of size, moisture or surface pollution, according to the companies manufacturing the machines.
In addition to a number of diamond case study examples (see The ore sorting effect below), tungsten and tin operations have also seen XRT machines work wonders.
For instance, Minsur’s San Rafael tin operation in Peru repaid the $24 million capital cost that came with buying a TOMRA 3,000 t/d XRT ore sorter within four months of installation.
The use of sensor-based XRT ore sorting converted uneconomic waste material into economic ore, according to TOMRA’s Robben, meaning material below the cutoff grade for the main plant, set at 0.9% Sn in 2014, was able to be treated with lower specific operating costs, thus bolstering reserves.
And, capacity of the San Rafael main plant freed up after the installation of the ore sorter, with nameplate rising to 3,600 t/d, compared with 2,950 t/d, Robben said.
Companies are also starting to realise the solutions can be applied at the front end of operations, too.
For instance, at a bulk sample trial in Western Australia for Novo Resources’ Karratha gold project, a combination of XRT and electromagnetic induction (EM) sensors were recently used to concentrate gold.
The nuggety mineralisation witnessed at Karratha had, up until this point, proven tricky to separate from the waste materials, but recent studies showed a combination of XRT and EM sensor-based ore sorting could do exactly this.
Novo said the XRT identifies rocks containing particles of high atomic mass such as gold, while EM identifies rocks that become electrically charged due to the presence of metallic particles.
At Vista Gold’s Mt Todd gold project in Australia the combination of XRT and laser-based ore sorting has shown a coarse fraction (+16 mm) high pressure grinding rolled product can efficiently be separated into waste and ore piles. The laser sensor, in this case, detects quartz in the ore after it has already gone through an XRT-based process.
Robben was quick to dispel the opinion that existing mines are reticent about using XRT ore sorting in fear of underfeeding their follow-on processing plants.
“As the application of sensor-based ore sorting often leads to a reduction of total cash costs, it turns resources into reserves and actually increases the life of the mine. Marginal resources are turned into reserves, for example, at Wolfram’s Mittersill mine in Austria, Coeur Alaska’s Kensington gold mine in the US or Minsur’s San Rafael tin mine in Peru.”
Kroukamp, who with Cronimet has overseen the installation and running of several XRT ore sorting solutions, explains why there may be a need to apply more than one sensor-based solution at mining projects in the future.
“Once you start going into finely disseminated tungsten ores, especially scheelite, you get these little spots throughout the rock…When you start going too small, the [XRT] machine fails as a result of detection. The signal coming back is negligible because of, let’s call it, background noise,” he said.
Robben said TOMRA is continuously developing the hardware and software to improve detection efficiency, and that it is tailoring its image processing to detect “specific mineralogical features”.
This detection issue may not be addressed by using more sensors, according to Guthrie.
“Multiple sensors can provide alternative ways to categorise the ore or waste, and potentially allow for more effective sorting, however the heterogeneity of the ore is what will be the deciding factor. It should also be noted that for low concentration elements or minerals, the sorters rely on associations with higher concentration minerals rather than attempting to detect the valuable component. This will rely on identifiable and consistent associations, which may not be possible with highly complex orebodies.”
The value proposition
Even if a mining company has a heterogeneous and distinguishable orebody suitable for ore sorting detection, it doesn’t necessarily mean ore sorting should be applied to it, according to Guthrie.
There are, broadly, three cases for improving project value through ore sorting, he said.
- Converting sub-grade waste (including previously rejected material) to ore that can be economic at modest recovery of values;
- Sorting of ore to increase grade and reduce subsequent processing and transport costs, and;
- Sorting of ore to bring forward cash flow.
Guthrie said: “The first case adds value to the project by conversion of waste to ore. On this basis, recovery is not critical and the value of the recovered material needs to cover processing costs plus margin. This can add significant value to a project where the processing capacity exceeds the ore supply rate and mine waste can be converted to ‘ore’ by simple low cost processing.”
Such applications are common in the mining space and where Cronimet normally starts its conversation with mining customers, according to Kroukamp.
Guthrie said the second case – sorting of ore to increase grade and reduce subsequent processing and transport costs – is generally only economically favourable if the processing and ore transport costs are high (in relation to head grade) and the recovery of values is also high (eg >95%) .
“This is particularly applicable to massive orebodies where there is little disseminated material but a component of the ore that is waste and has not been able to be separated in the mining process,” he said.
There are trials ongoing at existing operations that fit this mould, but Kroukamp said companies also need to factor in the existing infrastructure within this context.
“If you have a brownfields site, then you cannot underfeed these big mills and the beneficiation and flotation plants going after it by retrofitting sorters up front and cutting their feedstocks by, say, half. You end up with inefficiencies in the mill, which are already sunk costs,” he said.
Guthrie said recent work Ausenco has carried out indicates that “greater value can be realised through using a sorting system for upgrading of a low-grade material stream – for example low-grade stockpiles – to supplement higher-grade run of mine (ROM) ore that reports directly to the mill. In this case, the existing plant is kept at capacity and more metal can be processed”.
The third and last of Guthrie’s application cases requires that the treatment of the rejected material is deferred in the processing schedule to later in the mine life such that cash flow and net present value is increased.
“This requires that the cost of mining, sorting and ore storage is low and implemented without major capital and operating cost impacts,” he said.
Re-scheduling a mine and process plan to produce a higher-grade product at the outset would be very appealing to investors, but it has to be balanced with the value over the entire mine life.
One of TOMRA’s rivals in the ore sorting sector, STEINERT, has brought its magnetic and sensor-based sorting systems from the disposal and recycling sector into the mining industry.
STEINERT offers other sensors that can be combined with XRT to determine and sort individual chemical elements very precisely. Optical sorting and lasers, for example, are well suited to the detection of ores with different colours, such as copper oxide, or crystalline structures in quartz, and can be used alongside XRT.
One of the company’s multisensory sorting systems (KSS) recently made an impact at Braveheart Resources’ Alpine gold project near Nelson, British Columbia, Canada.
The purpose of the testing was to determine whether the STEINERT ore sorting system and methodology could be used to upgrade mineralised material originating from Alpine prior to trucking and processing.
The KSS machine in question incorporated a combination of XRT and laser sensor technology, with the XRT targeting the atomic density of the material and measuring the X-ray attenuation of each particle with a direct correlation on the mineral composition of the rock, and the laser sensor targeting the shape and brightness of a particle.
In preparation for testing of the ROM and composite samples, the -10 mm fines were removed by screening. During testing, the ROM material was upgraded from a feed grade of 14.7 g/t Au to 20.3 g/t Au with 92.8% gold recovery and 32.7% waste rejection. The composite material was upgraded from a feed grade of 25.4 g/t Au to 43.2 g/t Au with 81.3% recovery and 52.1% waste rejection.
Ian Berzins, Braveheart’s CEO, said the test results were encouraging and indicated the sorting of mineralised material at the Alpine mine could be an “important component of Braveheart’s overall beneficiation process”.
STEINERT said: “By using ore sorting equipment, ore concentrates can be created at very low cost in small or remote mining installations. The entire processing line can be planned in semi-mobile form and consists only of crushers, screens, belts and sorting machines.”
For all of the improvements ore sorting technology has brought to the mining industry, effective ore sorting on a bulk scale is still some way away.
Guthrie explained: “Although there have certainly been improvements in the sensing technology detection times and accuracy, there is still some work to be done to develop effective sorting systems that will be better than a quality grade control regime.
“The most important consideration is the heterogeneity of the ore and maintaining that heterogeneity through the ore handling train, ie blasting, digging, trucking and conveying. The heterogeneity will define the ‘batch size’ and the effectiveness of the system.
“The detection times of the conveyor mounted analysers can be a limitation when considering they are often mounted on fast moving conveyors and tens or hundreds of tonnes will have passed the sensor before an accurate reading is registered. This limits the selectivity of the bulk sorting systems and the batch size.
“The other limitation is in the bulk diversion systems. While there are a number of mechanisms available to achieve the diversion, the majority are slow acting and the fast acting gate style systems will experience high wear rates. There needs to be a bit more thought put into how to effectively divert tranches of waste away from the main mill feed system.”
The industry is gradually upping throughputs, however.
TOMRA, last year, released an upgrade to its Tertiary XRT, the COM XRT 2.0, which increased the ore sorting ante. The speed of the belt in the new design has been raised from 2.7 m/s to 3.5 m/s, while the more powerful X-ray system accommodates the sorting of larger-sized material due to better X-ray penetration, according to the company.
Ines Hartwig, Product Manager at TOMRA Sorting Mining, said: “The maximum size of the particles that the TOMRA COM XRT 2.0 can handle is between 100 mm and 125 mm, depending on the material, which also contributes significantly to throughput capacity.”
She noted that these higher levels of capacity are particularly valuable for larger mines as they reduce the number of machines required and, therefore, also decrease capital and operating expenditure.
The performance of this technology was proven at Ma’aden Phosphates’ $560 million processing plant at the Umm Wu’Al project in Saudi Arabia. TOMRA’s Sorting Solution division installed nine of its TOMRA COM XRT sorting units, each with an operational width of 2.4 m, to process a 1,850 t/h sorter feed at this facility.
The objective was to reduce the milling and flotation of silica in the plant process, using a dry technology. The TOMRA units achieved this by removing more than 90% of the chert in the +9 mm fraction, which makes up half of the plant feed, before the phosphate is fed to the milling and flotation circuit. This led to the removal of over 1.2 Mt/y of SiO2, which does not have to be crushed, ground and floated.
Metso, which has not yet developed a commercial product offering on this side of the market, thinks the industry should aim even higher.
“To make sorting viable for pre-concentration, it should be applied to bulk quantities of ore, such as on a loaded truck tray or a fully loaded conveyor belt,” the company said.
Olivier Guyot, SVP Mining Technology, for Metso, told IM that the company has been “interacting extensively with the ecosystem of companies in the ore sorting space for the last five years” and was “currently involved in several initiatives with partner companies”.
He added: “Most importantly, we are developing breakthrough proprietary technology to address the demand of high throughput accurate sorting.”
The Australian Commonwealth Scientific and Industrial Research Organisation (CSIRO) is working with industry players Advisian and RFC Ambrian on developing such a solution.
NextOre takes advantage of magnetic resonance technology to illuminate batches of ore with short pulses of radio waves, assessing ore grade and enabling high tonnage sorting at extremely high speeds, according to CSIRO.
RFC Ambrian Associate Director and acting CEO of NextOre, Chris Beal, said the sorter can pass material through at the rate of 5,000 t/h, and is able to distinguish “the wheat from the chaff ” in seconds.
The unit the company is working on is able to detect the magnetic resonance signatures of many minerals, including the most common economically significant iron-and copper-bearing minerals. It also has the ability to detect arsenic-bearing minerals often positively correlated with gold in sulphide ore deposits.
CSIRO Team Leader, Dr David Miljak, said the ore sorter is likely most suited to the copper mining industry as it can detect these minerals at extremely low grades.
He believes the processing costs for a big mining operation will drop radically for each pound of copper or ounce of gold produced – as much as 20% in some mines. “That’s according to the numbers we’ve put together. For an operation which spends half a billion dollars a year just to keep mining, that’s a big deal.”
CSIRO’s Dr Nick Cutmore, who heads up the project’s research team, said the sensing technology was previously evaluated at Newcrest Mining’s Ridgeway underground gold mine (part of the Cadia Valley operation) near Orange in New South Wales, Australia, with the tests demonstrating throughput capacity, accuracy and response times that improved significantly on competing sensing technologies.
Starting up stream
Ilpo Auranen, Chairman of IMA Engineering, thinks companies should start thinking about separating ore from waste at the earliest possible opportunity to improve their fortunes when mining begins.
IMA, which has a Sensor Solutions division, has carried out comprehensive studies on ore losses (OL) and waste rock dilution (WRD) across the industry and estimates mines are losing 10-30% of the ore value by sticking with the status quo.
The root causes, according to Auranen, include:
- Analysis and sampling averaging – Evidence of ore/waste heterogeneity is masked by common drill core and drill cuttings sampling and analysis practice where analysis technology and related costs is setting limits to sample size/length. This averages the ore grade on drill core, blastholes or RC hole length;
- Inaccurate sampling and analysis – Related to human error and defects, sampling and analysis technology. This results in errors in ore/waste distinction;
- Ore movement in blasting – Mixes ore and waste in contact zones;
- Mining method related OL/WRD – Mining waste rock blocks in ore leads to dilution;
- Internal waste in ore block and internal waste in waste blocks causes WRD/OL;
- Pit wall stability in open-pit mining and stope stability related issues in underground mining, and;
- Deleterious minerals/elements in ore reduce ore recovery and may turn ore blocks to waste.
IMA Engineering and Mine On-Line Service’s (MOLS) ore sorting study package includes an ore sorting simulation from drill cores (OSSCORE), drill cutting analysis from blasthole and RC hole drill cuttings, blast movement monitoring and laboratory studies using a Bulk Ore Sorting System (BOSS).
Auranen starts at the top: “OSSCORE analyses drill cores with MOLS’ Scanmobile lab using the XRF method at short length analysis intervals (20 cm on average), pinpointing the exact location of ore and revealing narrow veins of ore and deleterious elements.”
A comparison is then carried out with longer sample lengths of 3 m and 10 m, which represent typical block sizes in underground and surface mining.
“The OSSCORE is recommended to be done at an early stage in exploration and mining projects due to its positive impact on project financing,” Auranen concluded.
One of the recent success stories the company and MOLS had with OSSCORE was with Sotkamo Silver in Finland.
The two companies carried out an XRF ore sorting simulation using the Scanmobile lab at the company’s silver project, which has led to Sotkamo Silver updating its ore reserve and updating its mine plan to use an ore sorter in the ore preparation process.
It has also seen the planned total mining volume at the project increase from some 500,000 t/y to close to 1.8 Mt/y, boosting the maximum potential ore concentrating capacity to 600,000 t/y.
When it comes to drill cutting analysis from blasthole and RC hole drill cuttings, IMA has a sampling robot – the IMA Percussive Drill Sampler Analyser (PDSA) – to continuously sample and analyse the drill cuttings while drilling.
The average grade of the ore is given at every 20 cm to 100 cm hole length interval with the IMA PDSA providing a very dense map of the elements in the blasting bench and displaying accurate locations for ore and waste borders, Auranen said. .
Auranen said MOLS has a good understanding of the OL and WRD caused by ore and waste movement while blasting and offers Blast Movement Monitors, as the representative of Blast Movement Technologies in Scandinavia, to counter this. These monitors are installed in a blast pattern and move with the rock during a blast, providing valuable post-blast location information.
And, lastly, IMA has a pilot scale BOSS in its laboratory used for analysing customers’ rock and drill cutting samples and providing estimates of ore sorting effectiveness, Auranen said.
While the company has carried out this analysis in the lab, it is aware the best place to test is in the field.
“Lab samples are normally clean and dry and often screened to desired fractions and their volume is small,” Auranen said. “The ore at site often is dusty, muddy and wet and the volume is big.”
IMA, therefore, recommends studies to be made in the real mining environments using its sorting study service with leased IMA equipment.
For this, an IMA FCA (Fast Conveyor Analyzer) using X-ray fluorescence (XRF) sensors is usually installed on an existing belt conveyor(s) at the mine.
Auranen explained: “The FCA analyses ore continuously on a conveyor typically located after primary or secondary crushing giving average grades of the ROM bulk ore in short (five to 60 second) time frames. With this, the mine can get real information of their ore/waste ‘sortability’ and, at the same time, they see the ore grade and its variations in real time.”
The ore sorting effect
Some of the most impressive ore sorting case studies have come out of the diamond industry – a sector where the ability to pick up a plus-100 ct rough before it enters the crushing stage can be the difference between profit and loss.
DebTech, one of the De Beers family of companies, offers XRT ore sorting solutions to the market and recently won an order from Newfield Resources and the Tongo diamond project in Sierra Leone.
Working in collaboration with project house Paradigm Project Management, DebTech supplied a dry unit, CDX118CD dual wavelength X-ray sorter to the project.
The unit features an eight-channel photo multiplier detection system capable of identifying all types of diamonds including low luminescence, yellow and boart, the company said.
Gavin Alexander, Products Manager at DebTech, said: “The appeal of the technology is its efficient diamond recovery with minimum gangue material, even at high feed rates. These rates can range from 825 kg/h, with material sized between 1 mm and 2 mm, to 4.5 t/h, with material of 16 mm to 32 mm in size.”
There are manual and automated inlet chute gate options available, with an air ejector system that ensures no loss of valuable stones, DebTech said.
The diamond processing technology specialist has also been assisting South Africa’s largest diamond mine to optimise its plant throughput by re-concentrating the product from the DMS stage, with the recent delivery of an XRT sorter.
The XRT Coarse Concentrator sorting unit was delivered to De Beers’ Venetia mine, in Limpopo province, last year and commissioning took place during November.
The XRT technology has been specifically designed to provide an alternative to a high yielding DMS plant; especially important where the high yield causes plant bottlenecks further down the process.
DebTech Head, Gordon Taylor, said: “A valuable role that the XRT sorter can play is to re-concentrate the coarse and the middles DMS product. The significantly lower yield from the XRT sorters allows these size fraction to bypass the traditional recovery stage, thereby freeing up more capacity in the recovery plant to the treat the finer size fraction.”
Taylor added: “Not only is the XRT technology able to identify the diamonds, but the X-ray images captured also allow the machine to provide online carat estimates and stone count values in real time. This makes a big difference to mining and plant operations, especially in segments like marine diamond mining where the first installation of this XRT technology was done some five years ago.
These DebTech installations come on top of two high-throughput XRT ore sorters installed at Debswana’s Jwaneng diamond mine (Botswana) as key elements of its new Large Diamond Recovery pilot plant.
In Canada, Stornoway Diamonds installed a TOMRA XRT ore sorting plant last year as it looked to reduce the amount of waste that was finding its way through to the crushing and recovery circuits, while cutting down on costly diamond breakage.
The C$22 million ($16.8 million) plant was introduced after the primary jaw crusher and before the secondary cone crusher at its Renard mine, in Quebec, in the June quarter, and has continued to perform well for the company.
In the company’s September quarter results, Stornoway said: “The diamonds recovered since its introduction have exhibited lower levels of breakage than observed previously with comparable feed composition and, overall, breakage has been successfully maintained at sustainably low levels despite the high level of highly diluted, lower-grade material supplied to the plant during this period.
“In addition, higher than expected diamond recoveries have also been observed since the beginning of ore sorting, indicating that a more efficient liberation of diamonds is being achieved from the ore passed through to the main process plant.”
Drawn to magnets
Eriez Europe recently designed and manufactured 12 wet drum magnetic separators, three suspended electromagnets and four metal detectors for one of the largest polymetallic deposits in Kazakhstan.
Wet drum magnetic separators are typically used for heavy media applications, iron ore beneficiation, mineral processing and the reclamation or concentration of iron-bearing minerals, with Eriez manufacturing multiple diameters, widths and tank designs in order to accommodate the various requirements for the metallic ore processing industries.
Eriez said: “The principle of operation of the wet drum magnetic separator is always identical regardless of the tank design. The feed slurry containing both magnetic and non-magnetic fractions is directed into the ‘separation zone’. The magnetic material is then lifted out of the slurry and travels over alternating magnetic poles, washing out any entrapped non-magnetic particles before being discharged.”
The latest in Eriez’s range of wet drum magnetic separators includes a self-levelling design: “There are no discharge spigots to adjust with a self-levelling tank, maintaining constant tank level at any flow rate. The high-capacity tank design accommodates surges and fluctuations in the feed rate. The result is higher unit capacity and improved performance compared to conventional systems, saving customers on both capital equipment and running costs.”
For the project in Kazakhstan, Eriez advised installing suspended electromagnets in conjunction with metal detectors, ensuring both ferrous and non-ferrous metals are detected and recovered.
The suspended magnet removes the ferrous metal while the metal detector focuses on any ferrous metal missed by the magnet as well as non-ferrous metal such as manganese steel, aluminium, copper, brass and stainless steel. The electro magnet is mounted or suspended over a conveyor belt to remove large pieces of tramp iron, but the magnet can also be mounted over feeders or chutes.
These suspended electromagnets are available in manual and self-cleaning models. Manual cleaning magnets are recommended where only small amounts of tramp iron are present, while self-cleaning magnets employ a belt, running around the magnet face to provide continuous removal of collected tramp.
Suspended electromagnets can be installed across the belt at 90° to the product flow, in a diagonal orientation at approximately 45° to the material flow to aid with the removal of long rods and bars or above the discharge point of the conveyor. “They are vital to ensure the removal of damaging tramp metal from the process that can cause costly damage to conveyor belts at transfer points and to downstream equipment such as crushers,” Eriez said.
“The detection system operates by measuring the change in the electromagnetic signal of material being conveyed through the sensor area. Since the magnetic properties of a material are completely independent of conductivity, both magnetic and non-magnetic tramp metals are consistently detected.”
The equipment was despatched at the end of 2018 and Eriez will commission the equipment on site in Kazakhstan later this year, it said.
“Through one complete solution offered by Eriez, the site will be able to process a variety of final products and achieve the maximum return from their processing line,” Eriez concluded.