The Effect of Screening Panel Selection on Drain Rates and Media Recovery in a Dense Media Circuit

Media losses are a significant contributor to the operational cost in a dense media separation (DMS) circuit. Of these losses, up to 80 wt% can occur on the drain and rinse vibrating screens. Although these screens are an integral part of any DMS circuit, surprisingly little work can be found in open literature regarding the effect of aperture configuration on the drain rates (draining water and medium). To establish the effect of slot aperture size, experiments were conducted on a 0.6 x 1.2 m vibrating screen with polyurethane (PU) panels at slurry densities between 1.6 – 2.8 kg/L, and volumetric flow rates of 19 – 25 m3/h for both ferrosilicon and magnetite. Results showed that an increase in the volumetric flow rate led to an increase in drainage rate, moisture content and medium carryover. Once an optimum volumetric flow rate was exceeded, a further increase in flow rate resulted in a decrease in drainage rate with a sharp increase in moisture and medium bypass to the oversize stream. An increase in ferrosilicon slurry density led to a decrease in drainage rate with an increase in moisture and medium bypass with each of these rates increasing as the slurry density increased. For magnetite, a gradual decrease in drainage rate was observed with increase in slurry density and volumetric flowrate beyond the screen optimum point. It was found that an increase in aperture size from 0.63 x 8.8 mm to 0.8 x 8.8 mm to 1.0 x 13 mm resulted in an increase in ferrosilicon drainage rate of between 1.3 - 1.8 m3/m2/h depending on the flow rate and slurry density with a reduction in moisture and medium carryover of between 1.0 – 1.2 % and 14.1 – 20.2 kg/t/m respectively. For magnetite, the drainage rate increased between 1.0 - 1.4 m3/m2/h as the slot width increased from 0.63 x 8.8 mm to 0.8 x 8.8 mm while an increase in slot length from 0.63 x 8.8 to 0.63 x 12 mm only resulted in an increase in drainage rate of 0.1 - 0.5 m3/m2/h.