Reflux classification of South African coal.

Abstract

In mineral processing, the value of a particle is inexorably related to its specific gravity (Galvin et al., 2009). As a result, gravity concentration is widely employed in the beneficiation of valuable minerals from the associated waste mineral matter as separation of a feed into two or more fractions is accomplished according to particle density (Napier-Munn and Wills, 2006). Although gravity separation aims to separate particles primarily according to density, the size and shape of the particles also contributes to the separation achieved. Thus, the suppression of the effect of particle size in gravity separation is crucial in the mineral processing industry. There have been numerous developments in the field of gravity separation equipment that are able to selectively, and consistently, fractionate a feed according to density by manipulating hydrodynamic forces in different system configurations. This study investigated the gravity separation of fine coal samples with a relatively high ash content using a new and innovative technology, the Reflux Classifier. The novel design of the device incorporates a set of inclined plates attached to the top of a conventional fluidised bed. Thus, the device combines the uniform flow conditions of the liquid fluidised bed and the well-established throughput advantage of the lamella settler (Nguyentranlam and Galvin, 2001). The premise of the research entailed the concentration of small quantities of fine high ash South African coal (particles finer than 1000 μm) according to density through both batch and semi-continuous investigations. A laboratory scale reflux classifier with three distinct inclined sections (70° from the horizontal), consisting of 6, 8, 12 channels with perpendicular channel spacings of 6.50, 4.50 and 2.10 mm respectively was built and commissioned at UKZN. Batch test-work was conducted on each of the 3 configurations using fluidisation flowrates of 3, 6, 9 and 12 l/min. The investigation proved promising and moderate to high upgrade in the overflow product was achieved in the 8 and 12 channel configurations, with significant improvement at higher flowrates. With an operating fluidisation rate of 9 l/min, the upgrades ranged from 40% to 80% in the -1000 + 75 μm size range for all channel spacings tested. Moreover, in the size range comprising particles finer than 75 μm, for which gravity separation techniques are typically ineffective, a reduction in ash content from 60.71% to 36.81% was attained when using the narrowest channel spacing (12 channels with 2.10 mm channel spacing), which translated to an upgrade (reduction in ash content compared to the feed) of 39.72%. Furthermore, an upgrade in the product of up to 85% in the coarser size ranges (+600 μm) was realised. Overall, a reduction in feed ash content from roughly 60% to 36.59% was attained at a yield of 50.97% using 12 channels. These results encouraged semi-continuous tests on this configuration, which generated yields ranging from 57%-69% with a relatively low ash content, typically between 32%-40% (compared to a feed ash of roughly 60%), in the -106 + 75 μm size range. Additionally, consistently high upgrades were seen throughout the entire duration of the run for particles larger than 106 μm, approaching an upgrade of 85% at the coarsest particle size. An overall yield of 57.91% was achieved, with a reduction in overall ash from 56.69% in the feed to 33.11% in the product, which constituted an upgrade of approximately 42%. Particle re-suspension behaviour induced by high aspect ratios, which heavily promotes density driven separation, was also noted in the 12 channel configuration.