Dry beneficiation of South African coals using an upwards cyclic air flow

Abstract

Globally, coal retains its pivotal role as a significant energy resource and in manufacturing steel. This is due to its recognized abundance, cost-effectiveness, and reliability. However, the waterintensive nature of the current coal processing practices raises concerns about their practicality and sustainability, particularly in arid, frigid, and underdeveloped regions. Consequently, the existing coal processing practices may necessitate adaptation due to their unsustainable water usage, regardless of potential technical and economic feasibility improvements. Developing efficient dry beneficiation methods has emerged as an attractive area of research. This thesis presents a comprehensive review article that consolidates information regarding the principles and outcomes of commercially available and experimental dry coal beneficiation processes. These methods exhibit significant potential and, with careful consideration of potential limitations and further investigation, could find future application through the establishment of new large and small-scale plant installations, expansions, and processing of coal fines.

Among these methods, the SEP-AIR unit, developed by Gormash Export in Novosibirsk, Russia, exhibits promise in addressing specific challenges encountered by the South African coal processing industry. These challenges involve water scarcity and depleting reserves, which force the processing of finer particles (already being processed wet for many types of non-oxidized coal) and mining coal deposits in remote and underdeveloped regions. Traditional water-based mining methods may prove impractical for these conditions. The SEP-AIR system employs density separation using air as a medium to induce particle fluidization and subsequent stratification based on size, shape, and density. Diverging from most commercial dry, dense medium processes, the SEP-AIR employs suction force generated by a centrifugal fan and a specially designed separation nozzle to elevate particles and create a semifluidized bed with pseudo-fluid characteristics. The inventive separation nozzle design creates a vortex that captures sufficiently fluidized particles. The modular design of the SEP-AIR unit, coupled with its ability to handle dust, proves crucial. Additionally, the absence of water or additional dense medium materials eliminates the need for post-processing separation and drying, leading to decreased capital and operational expenses. While the separation efficiency of the SEP-AIR unit aligns with that of other dry processes, it may not reach the level achieved by wet processes. Nevertheless, it offers substantial throughput at high yield for particles ranging from 1 to 100 mm, including the typically discarded in dry cleaning (1-6 mm) for which studies are not as extensively reported.

Hence, this study aims to examine the potential of employing the SEP-AIR process to efficiently beneficiate smaller size fractions within Southern African coal types, encompassing the range of +3.2-13.2 mm. To achieve this objective, a laboratory-scale separation unit resembling the SEPAIR was constructed per manufacturer specifications. The unit was subsequently established and operationalized at the North-West University Potchefstroom Campus. A sequence of experiments was undertaken to investigate airflow characteristics, single-particle dynamics, multi-particle interactions, and assessments of coal separation efficiency. The findings from these experiments constitute the core content of this thesis.

Fluidization-based separation mechanisms are influenced by both the characteristics of airflow and particle interactions. Therefore, understanding the airflow dynamics and particle behaviour within the separation unit is crucial. To explore the former, this thesis includes an article detailing airflow simulation within the laboratory-scale SEP-AIR using Computational Fluid Dynamics (CFD) and subsequent validation experiments. The investigation includes three distinct airflow intensities: low, medium, and high. Comparison between modelled and experimental velocities demonstrates a notable degree of agreement. For the low flow system, the coefficient of determination (R2), root-mean-square-error (RMSE) and mean absolute error (MAE) values were determined as 0.95, 1.13, and 0.81, respectively. Similarly, the R2, RMSE, and MAE values for the medium flow system were found to be 0.93, 4.09, and 3.09, respectively. Lastly, the R2, RMSE, and MAE values were obtained for the high flow system as 0.94, 6.78, and 5.14, respectively.

A series of experiments were executed utilizing single particle tracers of 8 mm, with varying densities of 1.4, 1.6, 1.8 and 2.0 SG. These experiments aimed to discern the particle dynamics within the separation unit across different airflow settings. Qualitative observations were documented, focusing on the movement, fluidization, and retention of both single-particle tracers and multiple-particle tracers. From these observations, optimal parameters relating to airflow settings, residence time, and mass loading were determined to efficiently separate particles with distinct densities and sizes. It was noted that individual particles' behaviour aligned with the prevailing airflow conditions. Furthermore, the interactions among multiple particles in conjunction with the air during the fluidization process yielded the formation of a semi fluidized bed characterized by pseudo-fluidic traits. Consequently, lower airflow settings were found to be a prerequisite for achieving effective separation, compared to the conditions observed in singleparticle experiments. These characteristics were evident in the simulated particle tracks, corroborating the findings.

The thesis includes articles detailing investigations conducted to assess the potential of the SEPAIR for beneficiating smaller coal fractions in Southern Africa, specifically the +6.7-13.2 mm (referred to as 'pea-sized') and +3.2-6.7 mm (referred to as 'duff-sized') ranges. Within the former study, three coal samples-WB1, WB2 (from Waterberg coalfield in Limpopo, South Africa), and MOAT (from Moatize coalfield in Tete, Mozambique)-were examined, resulting in separation efficiencies (Ecart Probable Moyen - EPM) of 0.25, 0.05, and 0.04, respectively. Notably, high yields of 81%, 81%, and 82% were achieved for WB1, WB2, and MOAT, respectively. Improvements were observed in product ash values, with reductions from 36% to 26% for WB1, 20% to 10% for WB2, and 17% to 12% for MOAT. Calorific values increased from the feed's 22, 28, and 30 MJ/kg for WB1, WB2, and MOAT to 26 MJ/kg for WB1 and 32 MJ/kg for both WB2 and MOAT. Additionally, slight enhancements were observed in cut-point densities, transitioning from the feed's 1.63 SG to 1.53 SG in the product for WB1, 1.47 SG to 1.44 SG for WB2, and 1.44 SG to 1.40 SG for MOAT.

For the +3.2-6.7 mm particle size range, coal samples from the Waterberg and Moatize coalfields were investigated, yielding separation efficiencies (EPMs) of 0.15 and 0.05, respectively. The product ash value improved, reducing from 34.12% in the feed to 27.91% for Waterberg and from 16.52% in the feed to 13.12% for Moatize. Calorific values improved from 22.60 MJ/kg in the feed to 25.06 MJ/kg for the Waterberg and from 29.56 MJ/kg in the feed to 30.75 MJ/kg for the Moatize. The yield was reported at 75% and 95% for the Waterberg and Moatize, respectively.

Furthermore, slight enhancements were observed in the cut-point densities, with values changing from 1.61 SG in the feed to 1.55 SG in the product for Waterberg and from 1.44 SG in the feed to 1.43 SG in the product for Moatize. Furthermore, considering the varied coal qualities tested, different airflow settings were explored.

An Analysis of Variance (ANOVA) test revealed that coal feed quality, more specifically particle properties relating to size and density distributions, played a dominant role in separation efficiency and product quality. Moreover, the particle size separation, combined with the trends observed in particle misplacement, indicates that the separation mechanism is governed by particle weight and not density. Conversely, air speed notably affected product yield, with higher air speeds corresponding to increased yields. Notably, adjustments to fan settings (airspeed) had limited influence on product quality.

The efficacy of the SEP-AIR in beneficiating +3.2-13.2 mm Southern African coal fractions is evident through enhancements in ash and calorific values, achieved at substantial throughput and yield. While separation efficiencies achieved fall below those of wet processes employed for analogous size ranges, they remain consistent with other dry methods. The interplay between particle and airflow properties is important, affecting different aspects of the separation process. The characteristics of the fluidization and subsequent separation process are governed by particle properties such as size, shape, and density, as well as the interaction of these properties with one another, the air, and unit walls.