Among the many definitions and perspectives now used by different end users, one that is particularly appropriate is “the spatial characterization of key processing parameters throughout a mineral deposit,” Meadows explained, while pointing out that geometallurgy can be applied to both new project development and to existing operations. The integration of geological information with process design can help companies make more accurate decisions in a challenging business environment, he said, as well as helping to enhance production scheduling and the prediction of metallurgical results. And, in a financial environment where project costs are ever increasing, the application of geometallurgical concepts can help de-risk capital expenditure.

Organized by the U.K.'s Institute of Materials, Minerals and Mining (IOM3) and held at its London headquarters in early June, the conference attracted more than 50 participants from academia, industry, equipment producers and service suppliers. Geographically, delegates arrived from as far afield as Australia, Canada, Chile and India, while industry sectors represented included copper, uranium, iron ore and industrial minerals.

Aside from challenges in raising capital for new and expansion projects, current industry trends include falling ore grades, increasing power costs, and more stringent controls on water usage and availability, Meadows noted. In addition, there is upward pressure on plant throughputs—Cerro Verde in Peru began an expansion to 240,000 metric tons per day (mt/d) in 2012, and the plant at First Quantum's Minera Panama is being designed for a 300,000-mt/d throughput— and innovative technologies such as highpressure grinding rolls (HPGR) are receiving increasing interest as alternatives to traditional comminution routes.

Looking ahead, Meadows said, these trends are likely to continue, with orebodies not only becoming lower grade but also more complex and variable mineralogically. Add in factors such as shortages of experienced plant operators and higher wear rates on equipment because of the need to process higher volumes of harder ores, and there is a real risk that productivity will be affected.

With the mineralogy, particle size distribution and texture of a deposit governing the flowsheet design, understanding the mineral liberation size is directly tied into the mill sizing basis, he explained. Each orebody is unique, and has to be treated as such, with the development of computer technology in recent years having had a direct bearing on the automated analysis systems now available that can make mineralogical interpretation faster and more accurate.

To summarize, Meadows said, the implementation of a geometallurgical approach to a project demands a high level of communication between all of the disciplines involved—geologists, mining engineers and metallurgists. “A good geometallurgical plant de-risks a project,” he stated.

Modeling Project Value Chains
Dr. Peter Dowd and Stephen Coward looked at geometallurgical models and the quantification of uncertainty in mining project value chains. Geometallurgy will improve the profitability of mining projects by providing better understanding of the processing characteristics of all materials across the entire value chain, explained Dowd, professor of mining engineering at the University of Adelaide. Rock characteristics can be usefully classified as either primary or response variables, he added.

The authors’ view is that current geometallurgical modeling comprises: identifying the variables needed to understand critical process responses; finding ways to sample and measure these variables; and then developing techniques for the spatial estimation and simulation of these characteristics at the right scale into block models. What is currently missing, they believe, is the use of these models to understand the impact of variable and uncertain rock properties on process performance, project design, operational optimization and closure, and using that understanding to improve the value of the project.

Dowd illustrated the presentation with a case study to show the benefits of using multiple business cases in the production of cash flows that reflect the interaction of orebody variations and system constraints. Using the example of a surface-weathered polymetallic sulphide deposit, they used a net smelter return approach to select the optimum route for each ore block, either leaching or flotation, identifying the option that provided the highest monetary return on a per-ton basis. Instead of generating a single cash flow model, as would be the case using traditional concepts, their system used multiple simulations to generate a range of cash flows from which the optimum could be selected.

Geometallurgical models can provide a way to clarify project value and risks by specifically including rock property variability and uncertainty, they concluded. In addition, richer models of value and a better understanding of the consequences of variability and uncertainty allow project stakeholders to improve the decision-making process, with better decision-making leading to more resilient project configurations and more robust outcomes.

An Iron-ore Perspective
With iron-ore producers facing increased competition as well as the need to mine lower-grade resources, new challenges are emerging for which a geometallurgical approach can offer some responses. According to Professor Somnath Basu of the Indian Institute of Technology (ITT) Bombay, not only are grades lower (having fallen from above 60% in the past to below 40% now, he said), but grades are also more variable—a single steelworks may have different grade feedstock on a daily basis. In addition, gangue mineralogy is also changing, so the question becomes: how can processes such as sintering be made more tolerant to varying inputs?

Basu described research at IIT Bombay that is investigating how the mineralogical characteristics of iron ore affect the properties of the sinter produced from it, with the aim of optimizing the sintering process as a function of the ore being supplied. Different gangue characteristics can affect sinter parameters such as the reactivity, strength and reducibility, while a higher gangue content, or a gangue that reacts more slowly during sintering, can increase the production of slag and lead to higher energy costs as well.

The institute’s current research is looking at ways of better using leaner ores, as well as generating a lower slimes fraction during beneficiation. “Knowing the percentage of iron or silica in a sinter feed is not enough,” Basu said. “We need to know what is in it. The percentages may be the same, but different minerals react differently.

“What governs a certain microstructure, and what energy is needed to produce sinter of the right structure and quality? These are questions that a geometallurgical approach can help answer,” he concluded.

Linking the themes introduced by Dowd and Basu, Alastair Cornah from QG Consulting U.K. described how key linkages within the iron-ore value chain involve decision-making based on global and local spatial estimates of “primary” mineral deposit properties (such as in situ geochemical grades) and geometallurgical “responses” (such as screened, washed, or concentrated product yields and grades), which are derived from limited sample data. These spatial estimates are typically made using classical geostatistical techniques, which fundamentally assume that the attributes of interest behave additively.

Cornah stated that this assumption is generally appropriate for primary variables. However, since response variables may behave non-additively, their direct linear estimates (for example during resource estimation, mine planning and grade control) may be biased and should be treated with caution. Wherever possible, the primary variable(s) that underlie the response variable should be identified and estimated instead, with transfer function(s) between the primary and response variables being developed.

The implication of this in the iron-ore industry should mean enriching spatial estimation models of chemical grades (block and grade-control models) with spatial estimation models of mineral abundances and textures. These are the primary attributes of the mineral deposit that are independent of mining methods, plant configuration and material handling, but underlie key response variables such as product yields, product grades and, ultimately, the sinter performance of the products.

Cornah used a case study to demonstrate that, under certain circumstances, direct averaging of an iron-ore response variable can result in material biases. In this case, he said, direct averaging of iron recovery to concentrate into larger blocks results in some significant local errors. The mechanism used to test the additivity of that attribute could alternatively be used to test the impact of direct averaging of lumps and fines yields in resource estimates, grade control systems and mine planning, he suggested.

Copper Porphyry Variability
As FLSmidth’s Meadows noted in his keynote presentation, porphyry-hosted copper mines have the potential to be among the world’s highest-throughput operations. In addition, a number of big copper deposits have different mineralogy closer to surface and at depth, which can lead to challenges in terms of both comminution and mineral recovery.

Teck’s Quebrada Blanca in Chile is one example, and as the company’s Enrique Chait told the crowd at the meeting, feasibility studies on mining the hypogene deposit that underlies the currently producing supergene ore zone have been under way for some time. The transition to hypogene ore will mean a change from leaching to flotation, with increased crushing and grinding requirements.

Between 2008 and 2011, the company drilled 120,000 m for reserve definition, of which 13 holes were designed to produce information about the geotechnical properties of the orebody in different parts of the deposit—in particular looking at how these were affected in the various alteration zones present. Rock hardness was a key data target, not only for comminution but also for the prediction of drilling costs as the open pit is deepened.

The geometallurgical study defined 13 litho-alteration combinations that display different rock hardness characteristics. These could be grouped into hardness geometallurgical domains, such that it was found that green mica alteration is always related to harder rocks, potassic alteration with intermediate hardness, and phyllic and argillic alteration with softer rocks. Plotting out the occurrence of these within the orebody model provides a means for predicting future comminution and drilling performance, Chait suggested, while noting that the poor correlation found between the geotechnical properties and comminution parameters was probably a result of differences in measurement scale between the two.

Other papers presented at the meeting by Chilean authors looked at whether geochemical analysis can be used to identify geometallurgical parameters in porphyry copper deposits, and whether the inter-reactions between ore, gangue minerals and water during grinding can be predicted from the orebody geology. In the first instance, the authors from the Advanced Mining Technology Center at the University of Chile found that hydrothermal alteration zones can have a use in predicting the metallurgical behavior of rock and in improving the planning process based on the predominant mineralogy, while the second concluded— albeit on the basis of limited experimentation to date—that hydro-geochemical data can have predictive uses in this context.

An Azeri Example
One of the case studies presented at the meeting focused on Anglo Asian Mining’s Gedabek copper-gold porphyry mine in Azerbaijan. According to Dr. John Monhemius, the deposit was worked underground for copper between 1860 and 1920, then lay untouched until Anglo Asian revived operations there as an open-pit gold mine in 2009.

Two ore types are present, Monhemius said: an older VMS deposit, the “sulphide ore” and a younger, highly weathered porphyry stockwork—the “oxide ore”. The mine's current throughput of 1 million mt/y of ore gives an output of around 2,000 kg/y of gold, plus byproduct copper and silver. From the beginning, gold has been recovered by heap leaching, followed by resinin-pulp recovery, with a new agitation plant having been commissioned last year.

However, as Monhemius described, the high copper content in some of the ore can create problems in terms of cyanide costs, so the company has used geometallurgical methods to define a matrix of 27 different ore types in relation to their cyanide solubility. The matrix defines the ores in terms of being sulphide, transitional or oxide, and in terms of their copper and gold contents.

Secondary copper sulphide minerals such as bornite, chalcosite and covellite are the main culprits in this context, he explained, with Anglo Asian having included a SART (sulphidization-acidificationcyanide recycling-thickening) plant into its circuit to recover copper and silver from the pregnant solution as sulphides. A further problem is that copper and silver can passivate free gold particles by forming a surface coating on them, so the company has installed a Knelson concentrator after the grinding circuit to overcome this.

In all, Monhemius said, the use of geometallurgy has helped to classify the ores at Gedabek, with the plant operators then able to determine the best process route for them as they are mined.

Geometallurgical Aims
Developing early-stage predictors for likely processing behavior in a porphyry deposit was the theme of a presentation by Al Cropp and others. A senior consultant at MinAssist, Cropp described geometallurgy as being all about integrating geological and mineralogical understanding, and knowledge of rock properties across the entire mining value chain. To achieve this, cross-discipline understanding and communication is essential, he stated, with the overall aim being to reduce and manage project risk.

Using the example of a porphyry copper deposit, geological controls on rock properties can include alteration, mineralogy, the rock texture and hardness—all of which can influence processes such as comminution and flotation and, in the longer term, acid rock drainage (ARD). The type of alteration can influence comminution factors such as the hardness, the grain size distribution and the ease of liberation, as well as the generation of fines. Potassic alteration produces fewer fines during grinding, while rocks subject to argillic alteration are weaker, so generate more fines.

Again, advanced argillic alteration can have a major impact on flotation performance, whereas the effect of potassic alteration is low. On the other hand, the calcium and magnesium minerals associated with sericite alteration can affect the particle surface chemistry, with changes to the bubble performance and higher reagent usage.

Mike Hallewell, from SGS, noted that it has been using a geometallurgical approach since the mid-1990s, using the methodology in numerous studies and projects, from metallurgical sample selection to production forecasting contracts. SGS has pioneered the development of specific proxy tests and simulation software for geometallurgical analysis, he stated.

With the global industry faced with increasingly variable and lower-grade orebodies, the challenge is to develop different proxy tests that can estimate metal recovery with more complex metallurgical flowsheets. As a simple rule of thumb, Hallewell said, the more complex a flowsheet is, the more impractical and costly it is to provide the extensive suite of metallurgical testing needed to reduce risk. Hence a greater range of proxy tests is needed to generate data that can be included in a resource block model.

Hallewell explained that SGS recently signed an agreement with the Australian company, Corescan, to combine Corescan’s automated core logging technology with SGS’s on-site laboratory network and geometallurgical expertise. Automated core logging addresses the common problem that manual core logging is often inconsistent. With accurate logging, better correlation can be established with metallurgical performance, allowing metallurgical parameters to be distributed across the block model with increased confidence, he said.

Focusing on the theme of geometallurgy for comminution studies, comminution consultant Derek Barratt looked at factors such as the decisions that need to be made on the number and location of dedicated drill holes, and selecting the number of sample intervals when undertaking testwok at various stages of a mine design—from conceptual to basic engineering. He also looked at the selection of appropriate test protocols, the interpretation of test results, the selection and sizing of comminution equipment at nominal and design mill throughputs, and the optimization of the mine production plan through the input of comminution parameters into the mine block model.

Barratt described the differences in comminution behavior between heterogeneous and homogeneous ore types, with homogeneous ores showing consistent and predictable behavior over a range of sizes. They can easily be modeled by continuous functions when estimating breakage energy by size, he noted, whereas discontinuous models are better suited to heterogeneous ores with fractures that are weaker than the rock matrix.

As the interest in the London conference showed, geometallurgy is becoming increasingly important as understanding grows of the benefits that this approach can bring. Nonetheless, it is far from being mainstream yet, and there is still a considerable learning process to be undertaken.

And, as Meadows pointed out, mistakes can still be made. Poor sample selection and overblending of samples are just two of the common errors that FLSmidth encounters, he said, as is the analysis of too few chemical elements. Model building can also be defective: “Keep the model as simple as you can and only increase its complexity if is needed. A model must make mineralogical sense,” he stated.

Underestimating the complexity of ore and ore types can also lead to major plant problems, Meadows added, citing examples of both successful—geometallurgicalbased—and unsuccessful recent plant designs. Too often, poor communication between geologists, mine engineers and plant metallurgists results in poor plant performance. Applying a rigorous geometallurgical approach is one way of reducing this risk.