Coal Preparation » Fine Coal
Previous laboratory and pilot‐scale test work in ACARP Project C16040 (Galvin et al., 2009) had established the potential of REFLUX™ Classifiers (RC™) to beneficiate coal particles up to 4 mm in size, and even up to 8 mm in size, at separation efficiencies comparable to that obtained by dense medium cyclones (DMCs). In this new project (C19001) it was established that the performance observed previously is readily extended to full scale. This means that the REFLUX™ Classifier technology can now confidently be used to increase the performance and capacity of existing coal handling and preparation plants (CHPPs) with minimal capital expenditure. The primary screen size can be raised from around 1 mm up to 4 mm (or higher), which should release capacity by increasing the CHPP's throughput by at least 50 %. The +4.0 mm fraction can be sent to the existing DMC circuit which, with a lower fines loading, will be able to process a larger flow of coarse material and operate with higher medium recovery on the drainage screens. The ‐4.00 mm fraction can be sent directly to a REFLUX™ Classifier. The overflow product from the REFLUX™ Classifier can then be de‐watered at (for example) 0.5 mm, with the ‐0.5 mm material sent to a flotation circuit if required. The ability of the REFLUX™ Classifier to separate over a wide range of cut points also means that both it and the DMC circuit can be operated at the same incremental density, thus bringing further improvements to the overall plant yield when compared to intermediate circuits using technologies that struggle to achieve low density cut points.
A full‐scale test facility was constructed at the Rix's Creek CHPP to test the ability of a full‐scale REFLUX™ Classifier to beneficiate coal up to 4 mm top size and above. An RC™2020 unit was installed which had a 2 m diameter lower cylindrical section and 1 m long inclined channels with 14 mm lamella plate spacing. Feed to the unit was prepared by screening the ‐16 mm fraction of ROM coal from the CHPP. Formally the agreed target particle size was 3 mm wedge wire which is often regarded as equivalent to 4 mm. Initially screens with 4 mm square apertures were used, however these gave very little material above 2.8 mm. In later work, screens with larger 6 mm x 15 mm slot apertures were used, which gave a feed with significant levels of solids up to 5.6 mm in size.
The facility was operated over a period of 6 months during which time numerous sets of steady state feed, product and reject samples were collected. With a tight budget, compromises were necessary, making the operation of the circuit challenging. However, provided the auxiliary systems were working, the unit itself was stable over a wide range of conditions. The feed volumetric flowrates were varied from 210 m3/h to 368 m3/h with pulp densities in the range 16 to 37 wt.%. This gave solids rates in the range 50 t/h to 140 t/h, equivalent to flux loadings of 15 to 52 t/(m2 h) based on the cross‐sectional area of 3.141 m2 in the vertical section. Across this broad range the RC™ performance was robust. Typically the product +0.50 mm head ashes were less than 10 wt.% with reject ashes over 70 wt.%.
The average Ep value was 0.09 for the +0.50 mm composite size fractions, and this improves to 0.06 if only the +1.0 mm material is considered. The performance was consistent across the experiments, which cover control set points from 1200 to 1450 kg/m3 and feed solids rates from 21 to 52 t/(m2 h). The mass‐balanced density partition data for individual narrow size fractions shows that with decreasing particle size the sharpness of separation deteriorates and the density cut point drifts to higher densities. Both these trends are as expected for finer particles since their lower settling velocities makes them more difficult to separate and more likely to follow the majority of the water flow, which is to the overflow. Across all runs there was a steady decrease in density cut point with increasing particle size. In the range from 1.0 mm to 4.0 mm particle size the variation in density cut point was only of order 0.20 Sp. Gr. (except for Run 08 with about 0.30 Sp. Gr. variation). In most cases the REFLUX™ Classifier achieved Ep values less than 0.05 for size fractions larger than 1.0 mm.
Based on these limited sets of results, the REFLUX™ Classifier does appear to suffer from more drift in density cut point than DMCs, but the drift was still relatively minor. In contrast, the RC™ achieved significantly sharper separations (lower Ep values) for a given particle size than typical large diameter DMCs. Some simple circuit simulations were performed which showed that the combined impact of the larger cut point drift but lower Ep values of an RC™ compared to a DMC is that its overall performance will be similar to a DMC when treating ‐4 mm material. It must be recalled that the purpose of this project was never to try to demonstrate superior performance of REFLUX™ Classifiers compared to DMCs. Rather, the aim was to show that REFLUX™ Classifiers have comparable performance with DMCs in the ‐4 mm +0.5 mm size range (which has been demonstrated). This then means that inclusion of a REFLUX™ Classifier in the circuit will enable the aperture size on the CHPP primary screens to be increased, thus increasing plant capacity.