By New Acland Coal principal process engineer Michael Rodgers

REPURPOSING technology to suit coal mining is becoming a speciality for engineers at New Acland Coal (NAC). In our latest venture, NAC has taken technology originally designed to sort rubbish and given it a new application in coal sorting.

Using X-ray sorter technology originally developed for waste recycling industries, NAC engineers are able to eject reject material from raw coal before it enters the coal preparation plant (CPP).

Previously the technology has been used to separate mechanical components of wrecked automotive bodies from various components of domestic household garbage.

NAC CPP manager Robert Rashleigh said the technology had also been applied to ore sorting.

“While this technology is currently used in processing metalliferous ores like tungsten and nickel, our trial has shown that it can also process New Acland coal with good results,” he said.

“Should we move forward I believe this would be the first time the technology is installed in a coal mine in Australia and we are excited by the initial results of the trial.”

Pilot scale testing of the technology was carried out at STEINERT Australia in Bayswater, Melbourne, using coal sourced from NAC.

The process is relatively simple in practice. A stream of sized material is fed to the X-Ray sorter, spreads out on a wide horizontal conveyor and passes between the X-Ray source and a detector. Inside the sorter, the X-Ray beam passes upward through the moving stream and differentiates particles of coal and stone, based on an adjustable sorting algorithm.

As the stream discharges over the end of the conveyor, compressed air jets placed across the direction of flow are activated. Individual particles of target material are ejected out of the falling stream by the air impulse and separated from the flow. Either species can be ejected from the falling stream.

Dual energy X-Ray transmission scanners are used to measure the amount of X-Ray radiation absorbed within each particle at two different energy levels. Thereby, it is possible to calculate the average atomic density of each measured area by the controlling software, independent of the effect that particle thickness or size have on the absorption of X-Ray radiation.

Thus, the sorter can distinguish a coal particle from a stone particle with a very high degree of accuracy. However, for the sorter to be able to determine what is coal and what is reject material it had to learn the difference.

In 2015, NAC sent a selection of hand-picked samples to Melbourne that included discrete coal and stone, taken from the raw coal stockpile. These samples were placed in the source beam so that the machine could be ‘trained’ to recognise coal and stone.

The characteristics of the individual samples were read by the machine while stationary in the detector beam. From these readings, the machine’s separation algorithms were developed and these were used as the basis of separation when the pilot work was undertaken.

In December 2016 bulk raw coal samples, in the order of six tonnes each, were collected from the NAC run of mine (ROM) pad to be used in the pilot scale project. The bulk material was screened in the laboratory to produce test portions -150+50mm and -50 +25mm with 1.5 tonnes and 1.3 tonnes respectively being sent to Melbourne. Although the -25mm was not treated, 25mm is an arbitrary size and the X-Ray sorter can treat smaller sized material. The relatively large test amounts were used to give more reliable outcomes.

The prime target outcome of the pilot trial was to remove stone from generally poor quality ROM coal and at the same time ensure that no significant quantity of recoverable coal was lost.

To achieve this, a high cut point was targeted when upgrading CPP feed in order to minimise coal loss to reject.

The sorting algorithm derived from the 2015 NAC test samples was used in the pilot trial and the stone was ejected from the falling stream. The algorithm was adjusted to reject only the highest mineral content particles so that coal loss to reject would be minimal. The first drum of -50+25mm was fed through the machine and the separated streams visually examined.

While initial separation appeared to be effective, there were small amounts of stone in the product and clean coal in the rejects which appeared to be misreported material. These particles were hand-picked from their respective streams and re-fed individually to the machine.

In all cases, the particles reported correctly. This indicates that the sorter was detecting and ejecting the material type effectively, but that there was another physical mechanism causing the material to misplace. These hand-picked particles were retrieved and returned to the streams to which they initially reported.

For the -150+50mm test portion, the first drum was fed to the sorter using the existing sorting algorithm. As for the -50+25mm material, stone was ejected from the stream. The initial separation was visually unsatisfactory with significant amounts of coal and reject misreporting to their alternate streams.

In a similar manner to that of the first test sample, pieces of stone were removed from the product and individually re-fed to the sorter. In all cases, the material reported correctly. These hand-picked pieces were returned to the product stream.

The sorting algorithm was adjusted and the material was recombined and re-fed. An improved separation resulted, but there remained some coal-bearing material in the reject, some of it as inter-banded composite material.

Ideally, further adjustment to the sorting algorithm would have been made in order to reduce the amount of coal loss. However, we were seeing significant breakage taking place from repeated passage through the materials handling system. Because of this, no further adjustment was made and the remaining drums were sorted.

The product and reject streams were packaged to be returned to the laboratory and were subjected to Float/ Sink testing.

Yields and product ashes are presented in the table, along with partition parameters and calculated misplacement of coal and reject. The partition curves developed from the laboratory float/sink data show acceptably sharp separations, as denoted by the Ep values. The Ep value is a measure of departure of the centre portion of the partition curve from the vertical (a vertical curve would represent a theoretically perfect separation). The lower the Ep value, the sharper the separation.

Across the two size ranges tested, the pilot X-Ray sorter removed a combined total of 20.6% of the ROM tonnes as rock directly to reject. The targeted outcome was demonstrated and the overall pilot project objective has been successfully achieved with a number of notable benefits.

The most effective benefit from raw coal upgrade occurs when treating the poor quality seams.

There is also a potential metallurgical improvement within the separation process itself.

When treating poor quality (that is low product yield) coals, the quantity of reject material that has to discharge from the Dense Medium Cyclone (the major separation device in the plant and the one which treats the coarse material) is large and can lead to some of that material reporting to product when all of it should ideally go to reject.

Under such circumstances, this inclusion of reject in the product will increase the product’s ash content requiring the DMC’s separating density to be lowered in order to maintain product specification.

Lowering separation density will reduce yield. For poor quality coals, where we see X-Ray sorting being of most benefit, removal of stone beforehand can reduce or even eliminate this effect.

Eliminating stone from the CPP feed is equivalent to an increase in plant capacity, ie remove a tonne of stone from the feed and replace it with a tonne of coal.

Put simply, using this sorting technique is attractive to us because it means we can put a higher quality raw coal through the CPP, which in turn means a higher yield. Basically, by sending some of the coal through this sorter first, our CPP would be doing less work getting rid of reject, and more work processing product coal.

Right now we’re evaluating the outcomes from phase one of the trial. Ideally we would like to go to the next stage of the trial which could potentially see us bring a sorting unit onsite to New Acland. Using the new technology at the mine site will let us test it on a large scale and test how it interacts with the rest of our CPP operations.

The Bayswater pilot testing unit is similar in most respects to a commercial size unit, but has a 1-metre wide feed belt while the commercial units have 2-metre wide belts.

What we would like to do is put the technology to use in a full-scale, commercial operation. The pilot scale unit is restricted in the upper particle size it can treat, notionally 150mm; commercial sorters can treat particles in excess of 200mm. As well as this, the pilot testing unit’s feed system is not continuous, but treats batches of test material.

Recommendations from the equipment supplier suggest that particle size range should ideally not exceed 3:1. This requirement was factored in to the selection of the size ranges investigated.

Sorting occurs at a very rapid rate, based on high speed electronics and air impulse valves.