Mineral Processing and Extractive Metallurgy: Transactions of the Institutions of Mining and Metallurgy

The optimum plant capacity for a new mine is usually based on empirical studies or ‘rules of thumb’, subject to confirmation by detailed scheduling of the proposed mining operation. Historically the mining industry has a record of poor returns on investment and a high rate of project failure using these methods, with underperformance in grade being a common experience. Periods of high mineral prices tend to obscure this underlying problem. The assumption that ‘economies of scale’ will result from increasing throughput rates needs to be balanced by an awareness of the adverse effects of increasing the rate beyond a level that is supportable by the resource. For each scale of operation considered, it is a reality that for any intended head grade, at the associated intended cutoff grade, the actual head grade achieved will fall as the mining rate increases. This effect is known to people at operations but is not generally recognised in current ore reserve estimation methodology. Once recognised, this dependence of head grade on mining rate can be quantified and used to establish the economically optimum mining and processing rate for a new project. A simple analysis is proposed, which may be extended to detailed spreadsheet modelling for financial optimisation.

This paper presents guidelines for studies required for the development of mineral processing facilities from initial feasibility studies through to commissioning. Mining project schedule and cost overruns can often be attributed to inadequate metallurgical testwork, engineering and cost estimating leading up to commitment to the project. In some cases this may result from lack of understanding of, and commitment by the project proponent to, the requisite metallurgical and engineering studies during the development stages. Guidelines for metallurgical testwork, process development, engineering and estimating requirements for each stage of precommitment studies are described together with those for the engineering phase.

This paper discusses the methods used in the design of flotation plants, including benchscale batch and locked cycle tests and pilot plant trials. The methods used to establish appropriate flotation circuits as well as the interpretation of the test work data are also discussed in this paper. Careful and appropriate sample selection must be followed by equally carefully designed and executed flotation test work if a successful outcome is to be achieved. These steps are detailed and the correct use of each type of test and the information which can be obtained is described. Practical design considerations such as flotation time, type, number and size of flotation cells, and the means of froth transport are all important and appropriate test work can guide decisions on all these factors.

The Cadia Hill Gold Mine and the Hellyer copper, zinc and lead plant present different problems with widely different target grind sizes, liberation characteristics. The contrast is between Hellyer's sequential flotation circuit versus Cadia's ‘simple’ copper–gold flotation circuit. The common theme is that pyrite forms the principal floatable concentrate diluent. Hellyer ore contains finely disseminated chalcopyrite, sphalerite, galena and tetrahedrite. The flotation plant design was based on extensive benchscale test work (including locked cycle tests) on drill core and many months of operation of a 30 t h⁻¹ ‘pilot plant’ using the modified Cleveland Tin Mine processing plant. The Cadia concentrator was designed based on an extensive benchscale variability test work programme. Data from approximately a dozen locked cycle tests conducted on drill core and 2 weeks of continuous pilot plant trials of samples obtained from an adit into the orebody were used as the basis of design. The orebody is a low grade monzonite porphyry with disseminated chalcopyrite/bornite/pyrite (0·17%Cu, 0·73 g t⁻¹ Au). Methods used to establish the appropriated flotation circuits for the treatment of the Hellyer and Cadia ores, as well as the interpretation of the test work data for these particular operations, are discussed. The predicted versus actual operational plant is reviewed and the comparison was found to be satisfactory.

It is commonly believed that fine particles have low flotation recoveries. Indeed, size–recovery graphs often are ‘hill shaped’ with high recovery in the mid sizes and low recovery at the fine and coarse ends. Yet rather than being due to ‘overgrinding’, fines recovery may be low because the grind is not fine enough! The different flotation needs of fine and coarse particles have long been recognised, for example, in the old design for separate ‘sand and slimes’ circuits. Yet the principles may be overlooked in the desire for simpler circuits and larger equipment. Most plants now treat all particles together in one wide size distribution. Reagent conditions are set for the dominant coarser particles, so fines are starved of collector. Worse still, if there are significant mid sized composites, they often have to be rejected in cleaning to achieve target concentrate grade. Yet the conditions which reject mid size composites (collector starvation and high depressants) also reject fine liberated particles. In contrast to this common result, fines flotation can be excellent when flotation chemistry is tailored to them. After first recovering fast floating liberated particles, it is essential to adequately grind remaining composites to enable fines recovery. First, this allows lower depressant and higher collector additions since composites do not have to be rejected from concentrate. Second, finer grinding in appropriate equipment may be managed to narrow the size distribution to flotation, allowing reagent conditions to be set to suit the majority of particles. Third, grinding in an inert attritioning environment like an IsaMill increases fine flotation rates by producing clean mineral surfaces. An excellent case study is the installation of IsaMills in the Mount Isa lead zinc concentrator to grind lead and zinc rougher concentrate to P₈₀ of 12 μm and zinc cleaner tailings to 7 μm. Most plant losses had previously been in the sub 10 μm fraction, yet the finer grinding increased plant recovery by 5% lead and 10% zinc and also increased concentrate grades. Circulating loads dropped, reagent additions dropped in spite of the much higher particle surface area, and the plant became much more operable and responsive.

Sampling of mineral processing circuits for metallurgical control is typically done by compositing cuts of feed, tails and (if applicable) concentrate over eight or twelve hour periods. An exceptional problem exists in CIL/CIP circuits since the residence time in the leaching/adsorption tankage may be ∼20 h and the tails from any period are derived from material entering the circuit an average of 20 h or so earlier and thus do not relate to the head from the same sampling period. In addition, mixing in the series of tanks in a circuit also modifies the correspondence between input and output of the plant. A special purpose simulator has been therefore developed for a simple CIP/CIL plant in order to allow for the mixing and delay in the tankage and permit estimation of a head assay corresponding more truly to any tailings sample. In this way more accurate assessment of the day to day recoveries of gold and silver would be obtained. The model, programmed in Excel, assumes that the tanks are perfect mixers. By inputting the sequential head grades and considering that no leaching has occurred, the head grade which corresponded to the tailings for each period can be estimated. Allowance is made for variations in feedrate over the sampling period by computing the average flowrate of slurry and thus the residence time of each increment of the tailings. For a typical circuit even this simple model was able to greatly improve consistency of recovery data with variations being approximately halved by eliminating much of the anomalous variation arising from the lack of correspondence between head and tail particularly when significant changes in head grade are occurring. This has permitted genuine variations in metallurgy to be identified sooner and with greater confidence and thus operational shortcomings to be addressed more quickly. From the work reported here it has become clear that for meaningful recoveries to be obtained in a CIL/CIP plant on a daily or shorter basis then correction for delay and mixing in the tanks is essential.

This paper examines the effect of a number of ore and plant variables on recovery in a CIL/CIP plant. Data collected on samples composited over 12 h periods from approximately 12 months of operation were adjusted to compensate for the delays and mixing in the plant which produce a disconnect between feed and tailings in the same period. The adjustment method used was described in Part 1 of this paper. The adjusted data were shown to provide a meaningful calculation of recovery in the plant period by period with major reduction in variation in the calculated recovery. This opened the possibility of further analysis of the variations in recovery to examine the influence of other variables. The adjusted data were tested for validity in a number of ways thereby showing that deficiencies in the model and adjustment method were responsible for very little of the remaining variation in recovery. Solution loss, residence time and particle size were all shown to have a significant influence but the effect of changes in head grade was dominant. In all approximately half the remaining variance could be accounted for. While these results are specific to the plant studied they demonstrate that the adjustment technique is able to open the way to both more immediate interpretation of recovery data leading to better control and to more meaningful and detailed analysis of the plant behaviour opening the way to more effective optimisation.

One of the mining world's greatest challenges is to effectively use consumables (i.e. reagents and grinding media) to maximise metallurgical response, while minimising operating costs. This paper provides an example of a logical approach to achieve this. The current study, at Perilya's Broken Hill concentrator in the lead primary grinding/roughing circuits, examines the dependence of collector adsorption on E_h. The data demonstrate that the addition of a xanthate collector during grinding can improve galena flotation. Further, operating the ball mill under a superior grinding chemistry regime can lead to significant reductions in collector consumption while maintaining or improving the metallurgical results.

Mining companies are continuously searching for new technologies that can improve plant efficiency, by reducing operating costs and maximising metal recovery. Identification of how gold losses occur within an operating plant is crucial for determining appropriate operating strategies for improved metallurgical performance. Classical mineralogical and metallurgical techniques enable the operator to identify where losses occur in terms of particle size and mineral associations, however, how the gold is hosted within the mineral grains (for example, as solid solution or as fine microinclusions) is more difficult to determine. Secondary ion mass spectrometry (SIMS), a surface analytical technique is capable of detecting elements from ppm to ppb concentrations. Further, SIMS can be used to quantify gold in sulphide minerals, through comparison with implanted reference samples. Analysis of feed and tailing samples, from the processing circuit of the Kanowna Belle Gold Mine, in Western Australia using a combination of mineralogical, metallurgical and SIMS techniques was able to provide detailed information regarding the deportment of gold within this circuit. This analysis will be used to design and test process changes to improve gold recovery at a bench scale with possible implementation in the full scale plant. This paper provides details of the testwork methodology, and resultant laboratory study to improve gold recovery.