Testing was conducted previously to investigate the effects of cyclone geometry and flow rate on the occurrence of apparent bypassing when separating 150×25µm anthracite in a 25mm diameter cyclone. Apparent bypassing indicates that some fraction of the feed coal bypasses or short circuits directly to the clean coal and/or refuse streams, independent of the separating action in the cyclone. The result is a loss and/or contamination of the clean coal, which potentially limits the application of dense-medium cycloning in the cleaning of fine coal.

Experimental Approach

Two grades of magnetite were used for this series of tests. Type-I was an ultrafine magnetite with a top size of 15 µm, which was supplied by the Pea Ridge Iron Ore Company. Type-II was coarser (43% < 15µm) and was prepared by dry screening conventional (90% <45 µm) magnetite at 38 µm.

The cyclone testing was performed in a 25mm diameter stainless-steel cyclone. The unit was fabricated in such a way that each component was interchangeable to allow a wide array of geometries to be investigated.

Five sets of tests were conducted in this study. Initially, three series of tests were performed (set 1), with the size distribution of the magnetite being the only variable from one series to another. For each series, three inlet diameters were used, while keeping the remaining cyclone geometry the same (underflow diameter=6.5mm, overflow diameter=8mm, vortex- finder length=25.4mm, cylinder length=25.4mm, cone angle=10°).

The magnetite samples from sets 1 and 4 were weighed (slurry weight), dried, and weighed again (solid weight). The relative density of the slurries was calculated based on the percent solids. A size analysis of each of the samples from set 1 was done using a Microtrac SPA.

Results and Discussion

In the first set of tests, the classification of the 100PR magnetite was minimal. The classification that did occur was the greatest for the smallest inlet diameter as might be expected. Under these conditions, the cyclone acts primarily as a thickener rather than a classifier, and the differential (the difference between the relative density of the underflow and the overflow stream) is attributed to this effect. By negating the effects of classification, the resultant splits and differentials would be expected to be similar.

The next set of tests was run to investigate the correlation between magnetite classification and the partition curves at the same test conditions (using 100PR and 50PR magnetite). When processing coarser coal, it has been shown that the classification of the magnetite can impact separation efficiency. The partition curves for the 150×106µm and 75×53µm size fractions at each test condition for both magnetite grades.

For the 100PR magnetite, it can be seen that as the inlet diameter increased, the RD50 decreased for a given size fraction. The “best” separation (lowest Ep value) was obtained for test 2 using the 5mm inlet diameter. An increase in the inlet diameter generally resulted in an increase in the a-bypass, particularly for the 75×53µm size fraction.

It can be seen  that the relative density of the underflow stream and the relative density differential increased with the flow rate and also with an increase in the ratio of the overflow-to-underflow diameters . However, the relative density of the overflow stream was insensitive to a change in the flow rate for all the ratios. Changes in the ratio of the overflow-to-underflow diameter have been shown to influence the separation efficiency of coarse particles.