Coal Preparation » Fine Coal
There are significant coal losses in flotation due to oversize particles entering the flotation circuit, usually due to worn screens. Conventional flotation fails to recover a significant fraction of these coarse particles, typically larger than 0.5 mm. A robust plant design requires unit operations that can span a wide size range, allowing changes in the screen aperture due to wear to be fully accommodated. Similarly, a single process that can operate across the full size range of 0-2 mm could also be attractive, provided the process is simple to run, high in throughput, and efficient.
The performance of the Reflux Flotation Cell in recovering coal in the size range 0 to 2 mm was investigated at a laboratory scale. The original hypothesis was that coarse particles could be recovered by applying high gas fluxes to produce a concentrated fluidized bed of spherical bubbles at a volume fraction of order 0.5~0.6. Then, any coarse particle detachment should be followed immediately by re-attachment with nearby bubbles. While this re-attachment might be the case, the study showed clearly that coarse coal recovery was favoured by lower gas fluxes, below 0.5 cm/s. There was little impact on coarse particle recovery due to the introduction of high volumetric feed fluxes.
The project included two distinct investigations. The first was concerned with the flotation of coarse tracer particles added to the system one at a time under specific hydrodynamic conditions. Tracer particles up to a screened top-size of 2.0 mm were introduced to the RFC individually, as part of the feed water. These particles were strongly hydrophobic, having a relative density of 1.25-1.30, saturated by diesel. The number of particles that partitioned to the overflow and underflow were counted. The proportion of the particles reporting to the overflow then provided a measure of performance. The experiments showed that the coarse particle recovery was very high at low gas fluxes, especially flux values below 0.5 cm/s. The results were clearly sensitive to the gas flux, and broadly independent of the feed flux.
The second approach involved the flotation of industrial feeds at different pulp densities. Since the tracer particle study effectively involved only a two-phase system, the industrial feed provided an assessment as a function of increasing solids concentration. Thus, this work provided a measure of the effect of particle-particle interactions on coarse particle yield and hence combustible recovery. These experiments were benchmarked against the Float/Sink separation method at a relative density of 1.6. Where the benchmark combustible recovery was high at coarse sizes, the RFC combustible recovery was similarly high. In some experiments, the Float/Sink combustible recovery at a relative density of 1.6 fell to below 60% at the coarser sizes. Here the coal feed contains more near-density material, and is arguably much lower in hydrophobicity. Nevertheless, the RFC also performed well on these coals relative to the benchmark.
In general, the recovery was higher than for the Float/Sink result for finer sizers, and lower for the coarser particles, consistent with that observed for water based gravity separation. The cross-over point occurred at coarser sizes of nearly 2.0 mm when the Float/Sink combustible recoveries were very high. Where the Float/Sink combustible recoveries fell to low levels (less than 60% at an RD of 1.6) at coarser sizes, the cross-over occurred at finer sizes, closer to 1.0 mm. This simply reflected the rapid Float/Sink decline in recovery as the particle size increased.
Overall, single stage flotation at the optimum feed and gas flux was superior to two stage flotation at half the feed flux, with a high gas flux in stage 1 followed by the optimum gas flux in stage 2. In conclusion, the Reflux Flotation Cell was very effective in floating coarse particles provided the gas flux was below 0.5 cm/s. This gas flux also provides adequate carrying capacity for the coarser particles.