Dry beneficiation in a Reflux Classifier

Macpherson, Siubhan Ariana

Title: Dry beneficiation in a Reflux Classifier
Creator: Macpherson, Siubhan Ariana
Relation: University of Newcastle Research Higher Degree Thesis
Resource Type: thesis
Date: 2011
Description: Research Doctorate - Doctor of Philosophy (PhD)
Description: It is common practice in industry to beneficiate coal (and other minerals) to meet contractual obligations. This beneficiation increases the concentration of the valuable component in the product by removing the unwanted gangue or mineral matter in predominantly density based separations. Current beneficiation technology relies heavily on water-based separation devices that are highly efficient. However these require extensive water recovery operations and result in a wet product that has a reduced value and higher transport cost than an equivalent dry product. As water supplies in the mining areas of Australia become less reliable, there is renewed interest in beneficiation technologies that do not require water or water recovery operations and can produce a product of equivalent value to water-based beneficiation. This thesis considers dry beneficiation in a novel gas-solid fluidised bed, known as the Reflux Classifier, with the addition of a dense-medium and vibration. Prior to this work, the application of the water-based Reflux Classifier to the separation of particles on the basis of size and/or density had been extensively studied and the use of the pneumatic Reflux Classifier as a size separation device had also been investigated. The application of the pneumatic Reflux Classifier to the density based beneficiation of particles had not been previously considered. In the present study, the density based separation of particles in a laboratory scale pneumatic Reflux Classifier was considered. The apparatus was constructed of stainless steel and had a cross-sectional area of 20 mm × 100 mm with a plenum chamber and distributor below a 1 m vertical fluidised bed and a 2 m channel inclined at 70º to the horizontal above. A sand dense-medium was used to encourage density based separations and sand was continually fed to the bed in all experiments. The bed could be configured for batch separations with continuous overflow removal, or continuous separations with both continuous overflow and underflow removal. The whole apparatus could be vibrated to study the effect of the addition of vibration to the separations. The separations obtained in the Reflux Classifier were compared to separations in a 3 m vertical fluidised bed to determine the effect of the incline on the separation performance of the vessel. The majority of the separation experiments were conducted on batch quantities of plastic tracer particles in the size range -6.35 +1.0 mm with densities of 1300, 1600, 1800, 2100 and 2400 kg/m³. In batch experiments, tracer particles of a single density were placed on the distributor and the bed filled with silica sand (density 2600 kg/m³, -355 +90 μm). The gas and vibration were introduced simultaneously with a continuous feed of sand medium and the rate of elutriation of particles (both sand and tracer) from the bed was measured. The batch experiments investigated the effect of changing the vibration frequency and direction, including separations with no vibration, across the range of tracer particle densities. In continuous experiments, steady state between the feed and the continuous overflow and underflow of sand was established. The tracer particles in single density samples were fed into the bed in 10 minute intervals over a period of 40 minutes and their rate of exit through the overflow and underflow was measured. The continuous experiments investigated the effect of changing the overflow rate of particles at a constant feed rate and the effect of changing the gas rate. The sand fluidised bed acted as a dense-medium, or pseudo-fluid, enhancing the density separation of particles. Increasing the gas rate, or decreasing the vibration frequency both acted to increase the size of bubbles in the bed, which resulted in less homogeneity in the medium and thus less effective separations. As with all dense media operations, the effectiveness of the medium increased for larger tracer particles. An almost perfect density separation of particles greater than 4 mm in diameter was observed while for particles less than 2 mm in diameter the separation was more dependent on the convective velocity of the medium than the particle size. The rate of particle elutriation from the bed was first order dependent on particle concentration. The elutriation rate data for the batch experiments was modelled using a two-parameter model for dispersed plug flow to find the particle velocity and dispersion. The particle velocity was found to consist of a convective velocity due to the flowrate of the sand through the vessel that affected all particles equally, and a buoyant velocity dependent on particle size and density. This buoyant velocity could be predicted from an equation for particle terminal velocity in the Intermediate regime, with the density and viscosity of the medium dependent on the vibration conditions The density cut-point of the separations could be varied by changing the overflow rate, the vibration rate and the gas rate. Increasing the overflow rate or the gas rate led to a higher density cut-point while increasing the vibration frequency reduced the density cut-point. The separation efficiency decreased with decreasing particle size or increasing gas rate, but was largely unaffected by the overflow rate. The inclusion of the incline increased the separation efficiency by encouraging the refluxing of near density particles, as observed in the water-based Reflux Classifier. This refluxing is a unique characteristic of the inclined channel that was not observed in the vertical fluidised bed. The separation performance of the experimental apparatus was very good for the largest particles in comparison to water-based beneficiation devices but poor for particles below 2.0 mm. For particles in the size range -6.35 +4.0 mm the separation efficiency, measured by the Ep, was 0.04 for the batch separations and 0.06 for the continuous separations. This compares well with the 0.02-0.03 Ep of water-based dense-medium cyclones on the same size coal particles. The Ep for particles in the size range -2.0 +1.0 mm was 0.16 in the batch separations and 0.35 in the continuous separations which is significantly higher than the 0.07 Ep of the water-based Reflux Classifier and the 0.14 to 0.18 Ep of water-based spirals. From the tracer particle experiments the best conditions for a continuous coal separation were determined to be high frequency vibration, and minimal gas and overflow rates, with the Reflux Classifier performing better than a vertical fluidised bed. A separation of coal confirmed that the dense-medium pneumatic Reflux Classifier with vibration can perform good density based separations. The geometry used in this work restricted the processing capacity (tonnage) of the unit to approximately 1.5 t/m².h, with higher feed rates resulting in particles accumulating in the bed and causing blockages. A waterbased Reflux Classifier has a throughput capacity as high as 47 t/m².h. This thesis describes the first attempt to characterise the behaviour of a Reflux Classifier with sand dense-medium. The intention was to determine if such an apparatus could be used to separate particles on the basis of density and the potential for its application as an industrial process. The focus was on the changes in behaviour and the separation cut-point and efficiency with changing conditions. From this work it was concluded that the laboratory scale Reflux Classifier can separate particles on the basis of density with similar efficiency and cut-point to current industrial standards but the processing capacity is significantly below that of similar footprint water-based technologies.
Subject: beneficiation; Reflux Classifier; particle separartion; mining industry
Identifier: uon:7730
Identifier: http://hdl.handle.net/1959.13/808713
Language: eng
Full Text

Hits: 1486
Visitors: 2179
Downloads: 961

		Thumbnail	File	Description	Size	Format
View Details Download			ATTACHMENT01	Abstract	246 KB	Adobe Acrobat PDF	View Details Download
View Details Download			ATTACHMENT02	Thesis	8 MB	Adobe Acrobat PDF	View Details Download