Coal Preparation » Gravity Separation
This project provides the industry with comprehensive and accurate audit data on the Mineral Technologies LC3, LD7RC and LD7, and Multotec SX4 spirals for semi-soft coking coal (SSCC) and thermal coal applications.
Over sixty cuts that were taken over a >2 hrs for each test with considerable care exercised to ensure that the feed solids loading at each site was not excessive. The sample analysis included wet sizing using 14 sieves and then float sink testing of 17 densities for five size fractions for the feed, product, middlings and reject. Detailed partition curves for each size fraction were produced.
Standard spiral analysis is to combine two of the three outlet streams and to use the 'two-product' formula. This prevents separate analysis of the middlings separation, and subsequent detailed modelling of the spirals. A number of mass balancing methods were investigated and a robust method selected. All product and reject partition curves were almost perfectly 'smooth' and the standard Whiten equation gave an excellent fit for the product and reject splits, for virtually all size fractions. This indicates representative sampling and highly accurate laboratory analysis.
However, it is becoming widely accepted that an absolute minimum of eight densities, with four either side of the cutpoint, is required for accurate determination of the cutpoint and Ep. As the cutpoint is not known in advance, in practice at least 10 densities should be used. These rules were only refined after the conclusion of the test work, and the cutpoints were typically higher than expected, and so some reported Eps overestimate the actual values due to an inadequate number of data points about the cutpoint.
Being a single dataset for each, the results should not be used to select an optimal spiral brand and model. The Ep is heavily dependent on particle size, so the reported overall Eps cannot be directly compared as the underlying sizings at each site were quite different. Sites 2 and 4, in particular, had limited particles >1.0 mm, so their reported overall Ep was heavily weighted by the less efficient separation of the finer particles. Therefore, for modelling purposes, Eps and cutpoints for narrow size ranges should always be used. This is a general rule for all coal processing equipment. There continues to be an excessive focus on Ep in Australia that leads to the separation performance at high and low density often being ignored. The high and low density tails have a profound effect on yield. Therefore Organic Efficiency (OE) is arguably the single best separation efficiency parameter, and should be more widely used in Australia. OE has the advantage that its calculation does not depend on the determination of a partition curve as it is wholly dependent on the feed washability and calculated yield. The main disadvantage of OE is that it is affected by the feed washability curve (eg amount of near gravity-material), whereas Ep is largely independent of the feed washabilty.
A key result for all spiral models was that very high yields were achieved for the low density, coarse particles, and low yields of the >2.2 RD (ie low recovery of high ash material to product) for the -0.125 +0.075 mm size fraction. The yield of -0.125 +0.075 mm size fraction in some competing technologies would be almost 100%, regardless of the particle density. Some beneficiation occurred for the -0.075 +0.045 mm material, although the accuracy of the - 0.075 +0.045 mm reported partition curves were poor, as float-sink analysis of such fine material is extremely challenging.
No spirals were optimised for low cutpoint operation, with all fines circuits essentially being operated for 'maximum yield'. Therefore the secondary aim of determining the performance of modern spirals at low cutpoints (i.e. <1.60 RD) was not achieved. However, the spirals may be performing optimally for each site as there was little near-gravity material in the feeds and the actual separations were comfortably above the 'knee' of the yield-ash curves. Unsurprisingly, operation outside the manufacturer's recommendations leads to poor results. Data from ACARP project C19046 indicates that a cutpoint of <1.50 RD is possible from spirals. Future sampling of coking coal installations, at very low solids loading, and low volumetric flows, is required to quantify modern spiral performance at very low cutpoints.
Coal Spirals Handbook
The outcomes of this project include a Coal Spirals Handbook to provide coal preparation plant designers and operators with a concise source of information about coal spirals. It is organised into six sections:
- Section 1 - brief history of spirals applications.
- Section 2 - the operational aspects of spirals.
- Section 3 - the spiral selection and circuit design options for spirals.
- Section 4 - how the performance of a coal spiral can be determined.
- Section 5 - the current state of the art in spiral modelling.
- Section 6 - Appendices.
Spirals offer a number of advantages such as low operating and capital costs, operational simplicity, excellent tolerance to variation in feed conditions, low maintenance, and high reliability. Spirals also require far less complex water circuits and lower quality water than other technologies. Unlike most other known competing technologies, the spiral cutpoint does not monotonically increase as the particle size decreases, and so the variation in cutpoint versus size is less pronounced for a spiral. The wider adoption of spirals has been hindered in more recent years by the lack of accurate and reliable plant performance data. Most of the historical spirals data is for trough designs that have been superseded, or for cases where solids loading was excessive, or both.
The available data for modern spiral profiles (eg LD7, LD7RC, SX4, LC3 etc) is included in this document.