For those engineers at the sharp end, designing plants optimized in terms of currently available process options, the questions of greatest interest were probably: in which functions will high pressure grinding roll machines (HPGRs) and inert grinding mills offer significant advantages over conventional technology; and how can they best be integrated into new or expanded plant flowsheets. More specifically: can primary flotation circuits cope with feed coarse enough that we can dispense with ball milling? Can HPGRs be used for primary milling? And can stirred mills successfully be used for secondary and tertiary milling as well as in the regrind circuits of conventional flowsheets?

Metso and Sandvik on Optimized Crushing
A number of papers discussed optimization techniques specifically, two of which dealt with crushing. A team from Metso Minerals reviewed the development of in-house techniques for optimizing cone crusher design, while joint research presented by the Chalmers University of Technology Rock Processing Research group and Sandvik Mining & Construction concerned optimization of a crushing plant to achieve maximum profitability through high product quality and yield, combined with minimum waste output.

The Metso paper reviewed the company’s current 3D Discrete Element Method (DEM) modeling techniques, which have been developed to allow detailed study of the interactions between crusher design variables and actual rock mechanics. The firm has also developed a breakage model incorporating elements of Population Balance Modeling techniques to describe breakage as a function of the loads on the individual rocks. Combining this model with DEM has led to a multi-physics based comminution model that can be applied to crusher development. The authors said this model is sensitive to all aspects of crusher design including crushing machine parameters and ore characteristics, so it will facilitate the design of a “best fit crusher” for any given application flowsheet. It can also be used to fine-tune existing crushing installations, where most of the data required will be available or can be measured.

The authors did note some ongoing challenges relating, for example, to some hard-to-measure variables and to the validation process. But the relative values for all measurable parameters provided by the virtual crusher model have been shown to track extremely well with plant data over large operational ranges.

The Chalmers–Sandvik presentation discussed the parameters that influence the profitability of a crushed rock product and how a typical crushing plant can be optimized to maximize profitability by combining standard plant simulation techniques with quality and economic factors in a model. The simulations and optimization results are verified using measured data from an actual plant. The authors concluded that the empirical model can be used to accurately predict shape in different size fractions and noted that the optimization revealed that quite dramatic changes to the crusher settings can be used to combine high yield with good product quality. Further investigation of economic aspects of the optimization results is needed before the settings suggested by the optimization can be regarded as wholly reliable. The optimized plant increases gross profit by approximately 3% but more comprehensive studies of the overall process economics must be undertaken to determine whether a structural change should be made.

Tools for Operational Optimization
Two papers described and discussed the use of measured data for key mill performance factors to optimize mill control regimes.

Powell, now at the University of Queensland’s JK Mineral Research Center in Brisbane, Australia, together with van der Westhuizen and Mainza at the University of Cape Town (UCT) Center for Minerals Research in South Africa, reviewed the development in recent years of grindcurves that relate mill filling to performance indicators such as throughput, power draw and product size and showed how they can be used to achieve optimal mill operation.

Clermont and de Haas of Magotteaux International in Belgium first outlined current techniques for assessing milling efficiency, then described the company’s Sensomag system and how it can be used to monitor the internal dynamics of a grinding mill. This information can be used to optimize and control mill media filling in order to minimize production costs while maintaining grinding performance. Sensomag can be used on its own or linked to an automatic grinding ball loading system named Magoload.

The term grindcurve refers to milling response curves set up by a snapshot survey method developed by Powell and Mainza and used prior to conducting major survey campaigns. Van der Westhuizen et al combined full surveys with associated crash stops with snapshot surveys to derive a comprehensive set of curves for the South Deep SAG mill in South Africa. The Falmouth paper set out to first establish the value of developing grindcurves to help optimize mill control regimes. It is anticipated that, if a generic form of the response curves can be developed, the mill response can be recalibrated with snapshot surveys when the ore feed changes. Secondly, the authors discussed the calibration of a mill in practice.

They concluded that the relationships between mill filling and throughput, power and grind size provide a useful control optimization tool. Since the responses change with ore type they can be used as an indicator of feed variations and to provide the key information required to deduce the optimal control point. In turn this can be used to set the target for an optimizing control system and the control objective can be refined to maximize output within operating constraints. The authors propose further investigation to establish robust generic grindcurve relationships that can be applied to specific mills and different ore types with minimal calibration work, hopefully making this operating tool available to plant metallurgists and operators.

Meanwhile mill builder Magotteaux has developed its own snapshot tool. The Sensomag has been developed to continuously measure both ball load and pulp slurry positions inside a running mill—pulp density is an important parameter which influences the grinding efficiency. The main data are provided in terms of toe and shoulder angles. The principal element of the system is a polyurethane beam installed inside the mill that contains sensors able to measure ball and slurry presence. No complex interpretation of any indirect signal is required.

Ball load and pulp slurry detection are performed on a mill section at every revolution and the raw signals are sent through a wireless link to a central processing unit. The four media and pulp angles are then computed and transmitted online to the customer’s supervisory system via a standard OPC link or 4 to 20 mA electrical signals.

The authors discussed calibration and reported tests on several industrial grinding mills that validate the effectiveness of the system.

Summarizing, the authors said that the Sensomag is able to finely and independently follow pulp slurry and ball load level evolutions inside the mill and provide this key information, online, to the plant engineers. Understanding mill internal dynamics will definitely enable engineers to:
• Optimize liner design to obtain good relative movements of grinding media and pulp as well as avoid ball projection and liner breakage;
• Monitor liner wear and efficiency changes in order to optimize liner replacement;
• Improve grate discharge design to keep pulp level constant all through the mill length;
• Monitor interactions between pulp angles and media angles to detect load expansion due to pulp density change and to run the mill with the grinding zone properly saturated;
• Optimize and control mill media filling degree to reduce production costs while maintaining the same grinding performance.

Other promising optimization avenues are still to be explored.

A third tool, provided by Rosario and Hall, is an unbiased structure for comparing mill energy requirements. It is designed to remove the uncertainty concerning the net circuit energy savings calculated when comparing complete SAG and HPGR circuits, which results from the extra equipment needed for HPGR operation. Rosario is a senior metallurgist with AMEC Americas in Vancouver, Canada, and also a PhD student at the University of British Columbia (UBC), and Hall is an associate professor at the UBC Norman B. Keevil Institute of Mining Engineering. The structure was developed for HPGR vs. SAG mill complete circuits milling precious and base metals hard ores and was applied to the design of four complete circuits based on ore data from two sites.

The paper explained the design criteria, flowsheet development, modeling and simulation work (with JK SimMet) and equipment sizing, then presented and discussed the results for pure comminution energy drawn and complete circuit comminution energy power draw values. The authors also considered the significance of variables that their exercise did not take into account—ore variability over time, heating and ventilation costs (which are the subject of current research), clay content, availability and maintainability, steel usage cost in SAG milling and additional benefits for downstream processing attributable to HPGRs.

The authors concluded that the work produced an estimation of the real energy savings that can be potentially achieved in the treatment of similar precious and base metal hard ores and demonstrated that these saving are within the range of 11.7% and 18.4%. In addition, it was indicated that a smaller carbon footprint could also be achieved based on the elimination of SAG mill steel media usage and the authors plan future work regarding this matter. The limitations related to assumptions regarding clay contents, heating and ventilation requirements, ore hardness variability were unveiled and discussed. They also concluded that, independent of the magnitude of clay contents, heating and ventilation requirements and ore hardness variability, the HPGR potential benefits in energy and carbon footprint became so apparent that such an option should not be discarded at early stages of design studies at the present moment when climate change issues are highly valued.

One study presented at MEI’s Comminution ’08 conference concluded that selection of HPGR to replace a
conventional SAG milling circuit may yield significant savings in energy costs while reducing grinding
media consumption and operating costs.

Taking all of the studies into account, the authors concluded that the selection of HPGR to replace a conventional SAG milling circuit can yield significant savings in energy costs while reducing grinding media consumption and operating costs. HPGRs have a smaller footprint, equipment delivery schedules may be shorter, and they deliver a finer product size. The capital costs of installing an HPGR unit are generally higher than for the SAG mill option, typically by between 6% and 10%. Depending on the ore type and application, HPGRs can help achieve comminution circuit energy savings of up to 25% and reductions in operating costs of 10%–20%. And incorporating more than one HPGR in a circuit may create further energy savings. A further potential benefit of inter-particle crushing is improved recoveries from downstream processing as a result of microcracking occurring within the particles.

Mainstream Inert Grinding: Theory and Practice
One of several papers providing updates on progress with Anglo Platinum’s adoption of HPGR and mainstream inert grinding (MIG) with stirred mills asked the intriguing question: What if we already know? Authors Walstra and Curry of Xstrata Technology and Rule of Anglo Platinum argued that careful evaluation of available options for sequential treatments of an ore feed offers significant potential benefits resulting from higher concentrate grades. Such gains are relevant not only for South African mines facing electricity supply shortages but more widely as well. Having thoroughly explained the steps required for such an evaluation— which proceeds from detailed mineralogical analysis—they presented a case study of the MIG application at Anglo Platinum’s MPL South Concentrator (formerly the PPL Sandsloot Concentrator) in South Africa using an IsaMill with inert grinding media.

This concentrator has three sections, with A and B performing fully autogenous primary milling and the C section IsaMill regrinding the most competent ore pebbles from the FAG mills. Normally feed to the IsaMill is 250-300 mt/h with F80 at 75 to 100 μ and P80 being 53 μ. However, as a result of temporary equipment downtime upstream in the process, the IsaMill was also used for a period to grind significantly coarser feed (F80 210 μ), yielding a product P80 of 65 μ. Although the feed size was thus more than twice the design size, the net energy requirement was only 55% higher.

An IsaMill with the shell removed showing the grinding discs and product separator. IsaMill technology is being
adapted in mainstream grinding and regrinding applications in an attempt to achieve higher energy efficiency and
improved metallurgical performance.

Increasing the concentrate grade has potential knock-on advantages as well: lower haulage costs, reduced concentrate drying costs and more energy efficient smelting. The authors argue that the latter yields sufficient gains that greenfield projects can reduce smelter volumetric and power requirements by 50% and existing smelters may be able to double smelter metal ounces at little or no additional cost. In comparison with a conventional platinum process flowsheet, an energy efficient one with IsaMills instead of secondary ball milling and a total of 8 MW of IsaMills (with inert grinding media) installed in concentrate regrind applications should offer potential efficiency benefits including: a 35% decrease in direct electrical costs, a likely 50% reduction in total operating costs, a 25% overall reduction in secondary milling costs, a greater than 50% reduction in combined scavenger and cleaner flotation volumes needed, a much smaller overall plant footprint and, conceivably a lower carbon footprint too.