Dry beneficiation of -13mm South African coal using an air dense medium fluidization bed

Abstract

Coal remains an important energy-resource base worldwide as it is considered abundant, affordable and reliable (WEC, 2022). South Africa’s national energy-resource base is dominated by coal; with coal-fired power stations meeting about 83% of the maximum electricity generation in 2018 (DOE, 2019). South Africa’s need for coal-based energy is not likely to change in the coming 20 years, due to the lack of suitable energy alternatives. Coal is inherently heterogeneous by nature, and it requires liberation and separation of fractional impurities prior to consumption. The process is known as coal beneficiation and is achieved by a broad range of techniques and water treatment practices. Currently industrialized, wet beneficiation processes are known to produce high separation efficiencies however, as the name implies, it requires substantial amounts of water. Difficulties in handling, transport and storage as well as increased capital, operating and transportation cost result from excess water retained in coal (Choung et al., 2006). Furthermore, in water scarce coal producing and processing countries, including Mongolia, Australia and South Africa, suitable water supply is limited and as a result, wet coal beneficiation techniques is not ideal (Chen & Yang, 2003; Zhao et al., 2015). The industry is researching dry coal beneficiation methods in the hopes of finding a suitable alternative that is capable of producing good separation efficiencies. Dry beneficiation techniques, such as Air Dense Medium Fluidized Bed (ADMFB) system, can prove to be an effective solution to remove ash-forming mineral fractions from ROM coal to improve coal quality and mitigate application issues as well as minimize carbon footprint. This investigation has been divided into three main sections: 1. Dry beneficiation research studies. A comprehensive literature review compiled from numerous sources investigating ADMFB technology as a viable dry beneficiation technique. 2. Experimental evaluation and testing of a range of operating parameters to identify the best performance of -13.2mm South African coal in a batch ADMFB. 3. Particle segregation and fluidization behaviour in the ADMFB was explored by simulating -13mm South African coal with CFD Euler-Euler approach and by employing, Neptune, a commercial fluid dynamic software. A batch-wise ADMFB apparatus was used during the evaluation and testing stage, to investigate the performance of high- and low-density ROM particles combined with magnetite as dense medium (Geldart B particles). An experimental plan was set up to capture the effect of the operating parameters including, PSD, coal to dense medium ratio, addition of vibration and fluidization velocity on the performance of the ADMFB system. Product, middlings and discard ash content (%wt), calorific value (CV), density (g/cm3) as well as system air flow velocity (m/s) were used as process evaluation criteria. Local coal fired power station statistics such as ash content (%wt) and calorific value (CV) levels were also selected as reference point for the process evaluation. The quality of the coal used in this study was upgraded to varying degrees according to the set variables. During operation of the bed, a large fraction of high ash coal was found in the bottom layer of the bed, producing lower ash yield values to the top layer of the bed (13.0% to 22.4%) compared to ash yields of 82.2% to 74.8% in the bottom of the bed. Insignificant impacts on separation performance were observed for all +2.8mm fractions when adding dense medium (+300μm magnetite with d50 = 490μm) or applying vibration to the batch setup. Application of vibration to the bed lowered the minimum fluidization velocity of the bed, improved bed stability and assisted with particle striation. Little to no upgrading was seen for - 2.8mm coal particle size ranges. It was observed that the ash forming minerals in the -2.8mm feed coal were lower than that recorded for the +2.8mm, which lead to the conclusion that most of the valuable macerals have already been liberated and therefore the feed coal had a smaller fraction of mineral matter content. Separation of coal and magnetite at these size ranges proved to be challenging due to tiny magnetite particles attaching to the additional exposed surface area of the coal. Nevertheless, the product produced still adhered to demands and standards of the local thermal coal market and in some cases also meet metallurgical standards. This operation philosophy aids in downstream processing by eliminating magnetic separation and recycling of dense media. For the majority of the ADMFB experimental runs, the best cumulative products were achieved for runs that employed no DM in the absence of vibration. In a coal-magnetite bed, the magnetite is fluidized (in suspension) and the coal particles float or sink in accordance with density. The selection of the best DM (density, size and volume) is difficult as its properties, coal properties and operational parameters of the bed are interdependent when optimizing the performance of the ADMFB. The ADMFB density can be manipulated by fine tuning the DM (fluidization velocity, DM material density and DM particle size). Computational Fluid Dynamics (CFD) is a useful tool to study the dynamic mechanisms of fluidized beds and is applied in this study to better comprehend the selection of a suitable DM for effective particle segregation. From literature, it was reported that dense medium (DM) had a significant effect on the separation performance of the ADMFB, however during the experimental portion of this study this could not be substantiated. For the most part, it was found that the experimental beds performed better in the absence of DM. As a result, investigation into DM material density, DM fluidization velocity and DM to coal ratio was conducted by using the CFD simulation software, Neptune. A suitable DM will be selected by evaluating the performance of a binary mixture of dense medium and fine coal by employing CFD simulation. Throughout the study the CFD model was able to demonstrate key characteristics of an ADMFB as found by various researchers in the field. An improved understanding of the role of the DM during density-based segregation of particles was obtained. From the simulation it was found that the best segregation was obtained at a fluidization velocity of 23cm/s using a binary mixture DM consisting of 75% fine coal and 25% magnetite to separate high- and low-density coal particles.