A statistical ‘randomised block design’ full plant trial of magnetic conditioning of flotation feed was undertaken at Rio Tinto's Northparkes Mine in New South Wales. Magnetic conditioning of flotation feed has been shown to improve the flotation of <38 μm paramagnetic minerals consistent with selectively aggregating these paramagnetic minerals. The process was evaluated using shift composite samples that were analysed for copper and gold. Because the process targets <38 μm minerals, all samples were sized and the size-by-size recoveries were calculated for each sample. The size-by-size analysis targeted the effect on fine mineral, reduced the plant noise and shortened the test period required to get statistically significant results. An increase in gold and copper recovery was found in the <20 μm fraction to a high level of statistical significance. For the >20 and <38 μm size range, the copper recovery increased to a high level of significance, whereas the increase in gold recovery in this fraction was to a lower statistical significance. There was no statistical improvement in the >38 μm fraction for either metal. Approximately half the copper and gold losses at Northparkes are in the <20 μm size fraction. While the sulphide copper minerals chalcopyrite and bornite are known to be paramagnetic; gold is not paramagnetic. Gold's response to magnetic conditioning, however, is consistent with a number of literature references where, depending on the gold's mineralogical disposition, gold has been shown to respond to magnetic treatment.
The tracking of ore through process hold-ups such as stockpiles has been traditionally been difficult to achieve. The inability to pinpoint the exact source in the mine of material that is feeding a processing plant has led to inefficient mine-mill reconciliation processes. In addition, a lack of understanding of the behaviour of ore in process hold-ups such as bins and stockpiles often leads to poor predictions of plant performance and to poor reconciliation of this to the orebody.This paper reports on two short case studies. One case study addresses tracking ore from various parts of two mines to the concentrator. The other follows synthetic particles through the coarse ore stockpile. Both studies use radio-frequency identification (RFID) tracer technology. The paper concludes by discussing general strategies for implementation of such concepts into mine-mill metal accounting and reconciliation systems. As shown in this paper, reliable ore tracking systems will allow much more effective mine-mill reconciliations as well as the ability to model process hold-ups with regard to size by residence time distributions. This should, in turn, allow metallurgists to have a better understanding of the size and grade of material that will report to the plant at different process hold-up (eg stockpile) conditions.