Published Feb 24, 2015

Eco-efficient and cost-effective process design for magnetite iron ore

Currently, the mining industry is facing several issues related to energy consumption so the optimized use of energy is an ever-increasing need. Mining – and especially minerals processing – routes for different ores (base metals, iron ore, bauxite, platinum, etc.) vary significantly, and the energy requirements and the opportunities for reduced energy consumption are also different. Iron ore has a special place in the global mining industry, judging by the volumes of ore processed and the energy usage.

Authors: Alex Jankovic, PhD, Metso; Walter Valery, PhD, Metso; Roberto Valle, MBA, Metso

Figure 1. Energy and greenhouse gas impacts of mining and mineral processing operations (source: Norgate, T y Haque, N. 2010).It is very well known that energy production also implies emission of CO2, as shown in Figure 1 ( The information indicates that almost 50% of the total CO2 emissions are generated by the comminution processes (crushing and grinding operations). For this reason, it is crucial to innovate through new technologies right from the conceptual phase to determine the best process route or circuit configuration.

Moreover, some countries are already imposing taxes on the emissions of greenhouse gases which is certain to have a negative effect on the process operating costs. Figure 1 shows the amount of CO2 emissions in each of the unit operations relating to mining operations and mineral processing.

The majority of steel production is supported by iron ore sourced from high-grade hematite deposits, although a significant fraction comes from magnetite deposits. Compared to direct shipping hematite ores mined from the upper regolith, magnetite deposits require significant beneficiation, which typically involves grinding to a particle size where magnetite is liberated from its silicate matrix. Many banded iron formation deposits are very fine grained, often requiring a final concentrate grind size P80 of 25-35 μm (see liberation curve of magnetite in Figure 2). The amount of energy required to produce a magnetite product suitable for sale as pellet plant feed from these deposits is an order of magnitude higher than an equivalent direct shipping lump (< 32 mm > 6 mm) and fines (< 6 mm) hematite project.

Figure 2. Liberation curve for the study case

The cost associated with high-capacity processing of a hard, fine-grained, silica-rich magnetite ore is presented in this paper, with the emphasis on comminution circuit options. The objective is to evaluate several options involving different grinding technologies with respect to energy consumption, operating cost and capital cost. Therefore, a typical conceptual or scoping level assessment methodology used by engineering companies was applied.


The key objectives of the evaluation are as follows:

  • Assessment of different eco-efficient comminution process circuits to treat magnetite iron ore.
  • Determination of the process operating costs and the capital costs for each process route.
  • Comparison of the benefits of the different process routes from an economics point of view.
  • Innovation and development of efficient technologies that enable the economic project viability.

Development and data collection for magnetite ore grinding

Various magnetite ore grinding flowsheets have been implemented in the past, including:

  • Conventional three- (and four-) stage crushing followed by primary and secondary milling.
  • Primary crushing followed by wet SAG or AG milling and ball or pebble milling.
  • Air-swept AG milling (for coarse grind).

Historically, the lowest operating cost for fine-grained ores was achieved by multi-stage, fully autogenous grinding (Koivistoinen et al, 1989) with integrated magnetic separation steps between the stages. The major benefit of fully autogenous grinding is the elimination of steel grinding media costs and the need to discriminate between steel and magnetite in coarse magnetic separation ahead of pebble crushing. The separation step between grinding stages progressively reduces the amount of material to be ground and, in many cases, reduces the abrasive properties of the concentrate.

Some of the best known magnetite companies using autogenous milling are the subsidiaries of Cleveland-Cliffs Inc. in North America. The original autogenous milling circuit, consisting of an AG mill followed by cobber magnetic separation of pebbles, pebble milling of the magnetic concentrate, a finisher magnetic separation stage and silica flotation, was installed at Empire Mines in 1963 (Weiss, 1985). There have been three expansions since and, in the 1990s, Empire Mines had a total of 24 individual concentration lines and a total plant capacity of 8 Mtpa of pellets. The target grind size of the circuit varies between the 90-95 percent minus 500 mesh (32 μm) depending on the ore and operating conditions (Rajala et al., 2007). For this specific case, Figure 2 shows the liberation curve for the magnetite ore.

New Technologies

Significant reductions in the costs associated with grinding were achieved over the first 80-90 years of the last century by increasing the size and improving the design of the crushers and mills; however, there was no major breakthrough in improving the energy efficiency of the comminution process.

The principles of particle breakage in crushing and grinding equipment remained mainly unchanged with the energy efficiency of the comminution process reducing as the product size decreases.

Only in the last 20 years were the more energy-efficient technologies successfully implemented at an industrial scale, including high-pressure grinding rolls (HPGR) for fine crushing (Dunne, 2006) and stirred milling for fine grinding (Gao et al., 2003). The application of more efficient grinding technologies has provided opportunities to further reduce the operating costs associated with grinding. At Empire Mines, an HPGR was installed for processing crushed pebbles, and its introduction resulted in a primary AG mill throughput increase of the order of 20 percent (Dowling et al., 2001). The application of Vertimill® fine grinding technology at Hibbing Taconite Company enabled processing of lower grade ores and increased the concentrate production (Pforr, 2001).

A sharp increase in the application of HPGR and stirred mill technologies is recorded in the last decade, driven by the benefits of increased energy efficiency and supported by improvements in equipment reliability. The potential for the reduction of energy consumption of the order of 30-45 percent was suggested to be possible (Valery and Jankovic, 2002), although significantly lower reductions, 9-13 percent, were reported after detailed engineering studies for two large copper projects (Seidel et al. 2006). This clearly indicates that benefits from new energy-efficient technologies are case specific and the intention of this paper is to show the potential for the magnetite ore processing.

Table 1. Ore design parameters

Study Options

A study into the options for a 10 Mtpa ore processing plant for a hard, fine-grained, silica-rich magnetite ore was carried out, with the emphasis on comminution circuit options. The concentrator was assumed to be located within 100 km of a port suitable for facilitating equipment delivery. It was assumed that there were no restrictions on spatial layout and that the process facility would be built on ground of a sound geotechnical character. Any subsequent differences in tailings disposal, water recovery and their associated operating requirements and costs were not considered.

A set of ore comminution properties used as the basis for this hypothetical study is provided in Table 1.

The magnetite concentrate weight recovery, SG, Ai, iron and silica content were based on the relationships presented on the left.

The fine-grained nature of this hypothetical ore results in a relatively late release liberation curve. This fundamental property of a magnetite ore is generally one of the major drivers of flowsheet design and, therefore, flowsheet option generation.

Four circuit options were selected for comparison (McNab et al, 2009) with the following acronyms used to identify the primary unit process within each:

COS – coarse ore stockpile; SC – secondary crush; HPGR – high-pressure grinding roll; AGC – autogenous mill in closed circuit with cyclones and pebble crusher; RMS – rougher magnetic separation; CMS – cleaner magnetic separation; CMS2 - second cleaner magnetic separation; PM – pebble mill; PC – primary crusher; SM – stirred mill; and TSF – tailings storage facility.


Primary crushing – AG milling in closed circuit with hydrocyclones and pebble crushing – rougher magnetic separation – pebble milling – cleaner magnetic separation.

Option 1 resembles the well-known fully autogenous LKAB and Cleveland Cliffs style, low operating cost operations. The absence of steel grinding media is the major basis for the low operating cost. Pebble mill control and pebble transport and handling requirements add complexity to the design and operation.



Primary crushing – AG milling in closed circuit with hydrocyclones and pebble crushing – rougher magnetic separation – ball milling – cleaner magnetic separation – tertiary milling using stirred mills – second cleaner magnetic separation.

Option 2 has an additional grinding and magnetic separation stage compared to Option 1 and is considered to be simple for design and operation. The final milling stage is carried out using energy-efficient stirred mills. Steel grinding media usage significantly increases the operating cost.


Primary crushing – closed circuit secondary crushing – closed circuit HPGR – rougher magnetic separation – ball milling – first cleaner magnetic separation – tertiary milling using stirred mills – second cleaner magnetic separation.

In Option 3, secondary crushing and HPGR effectively replace AG milling with pebble crushing. The application of HPGR, stirred milling and an additional magnetic separation stage reduces the power requirements compared to Options 1 and 2.


Primary crushing – secondary crushing – screening – Open HPGR – coarse pebble milling – rougher magnetic separation – fine pebble milling – first cleaner magnetic separation – tertiary milling using autogenous stirred mills – second cleaner magnetic separation.

Option 4 is an attempt to design a circuit with the lowest operating cost through increased grinding energy efficiency using three stages of magnetic separation, traditional autogenous milling, HPGR and stirred milling technology. In this conceptual flowsheet, steel grinding media is eliminated. Circuit complexity is partially reduced by open secondary crushing, HPGR grinding and stirred milling operation, although recovery, storage and control of three separate-sized media streams are introduced.

Discussion of results energy consumption

Figure 3. Energy comparisonWith the exception of the primary crushing module, which is consistent between options, estimates were developed for the total power drawn in the comminution, classification and magnetic separation areas of each circuit. Energy consumed by material transport machinery related to pumping between areas was not considered at this level of the study. A summary of the comparison of unit circuit energy for each option is shown in Figure 3.

A significant energy reduction is predicted for Options 3 and 4, which include HPGR and stirred milling. Some 33 percent of additional energy separates the most energy-efficient option (Option 4) from the least efficient, the two-stage AGC Pebble circuit (Option 1). Note that part of the energy reduction is also due to the fact that the process uses unit operations that are better suited to each stage of grinding, i.e. stirred mills are much for efficient for fine grinding than tumbling mills. It can also be attributed to the fact that Options 3 and 4 have an additional separation step at a coarse grind, which reduces the amount of material for fine grinding.

According to Seidel et al. (2006), the basic comminution energy requirement for the Boddington HPGR circuit option was 14 percent lower than the SAG option; however, the overall energy requirement, including conveying, screening etc, was reduced to 9 percent. The Boddington copper gold ore is of similar rock competency to that selected for this study and thus provides a good contrast between comminution processes designed to liberate minerals for flotation, in which the whole ore is ground to fine size, and the comminution process with the staged rejection of silicates. In the latter case, the energy consumption difference between flowsheet options can be significantly higher.

Process operating cost (OPEX)

Table 2. Key operating cost inputsA fairly detailed approach was taken in terms of the development of operating costs for each option. Consumption rates for power, wear and other consumables were considered for each process flowsheet. Maintenance and materials, as well as labor, were also considered. The scope covered included the process from the COS reclaim feeders to either the final magnetic separator concentrate discharge or the magnetic separator tailings discharge. As such, no concentrate or tailings handling, filtration or storage costs were considered. For simplicity, some minor operating costs, such as metallurgical testwork and analysis, which is considered common to all options, have been omitted.

Unit costs for power, grinding media, wear consumables and labor were referenced from average values within the GRD Minproc database for similar-sized and located projects. A factoring approach from the direct capital cost was used to develop cost estimates for maintenance materials. Key assumptions are listed in Table 2. All costs are estimated in Australian dollars and are presented as 1st quarter 2009 costs.

Carbon Tax

A carbon tax is expected to be introduced in the near future and would add a significant cost to all operations. For this exercise, a simplified estimate of the effect of a carbon tax is considered. It was assumed that the carbon tax would be applied to total circuit energy and steel consumption relating to media and comminution equipment wear liners. The following criteria were applied for the carbon tax estimate: CO2 emission, 5 t per 1 t of steel media (Price et al, 2002), CO2 emission, 1.0 kg per kWh of electricity, CO2 tax, $23 per t of CO2 (Australian Government, 2008).

Figure 4. Operating cost comparison

The estimates as summarized below are estimated to have an accuracy of ±35%. Unit cost breakdowns are presented and shown graphically in Figure 4.

Option 1: 6.17 $/t, Option 2: 6.42 $/t, Option 3: 6.66 $/t, Option 4: 5.38 $/t

The most significant operating cost (OPEX) variables between options are those relating to power, media and liner consumption. The two options including AG mill circuits have between 27 and 32 percent higher power consumption costs relative to Option 4, which utilizes the more energy-efficient autogenous grinding technologies.

Grinding media and wear lining costs range between 0.41 $/t and 1.82 $/t. Option 3 has much higher media and wear lining costs because two ball mills of 8.8 MW installed power each are required to grind 8 Mtpa of RMS concentrate from P80 2.3 mm to P80 75 μm. The overall OPEX for Option 3 is the highest due to the high costs of media and liner wear.

Table 3. Carbon emissions and carbon tax summary

Table 3 shows a summary of calculations related to the carbon emission and carbon tax effect on OPEX. It can be observed that the introduction of carbon tax at 23 $/t would increase OPEX to the order of 9-11 percent. The majority of carbon emission is from electrical energy consumption, while the indirect contribution from steel consumption is dominated by grinding media and is of the order of 5-16 percent for the options that utilize ball milling (Option 2 and 3).

Capital cost (CAPEX)

The scope of the estimates follows the Work Breakdown Structure developed specifically for the study

and considers each flowsheet from the COS reclaim feeders to either the final magnetic separator concentrate discharge or the magnetic separator tailings discharge.

The CAPEX estimate is developed based on the premise that the process is located inland in West Australia. All costs are estimated in Australian dollars and are presented as 1st quarter 2009 costs. They are estimated to have an accuracy of ±35%, which is commensurate with the accuracy requirements for a high-level options study of this nature. The details of the cost estimate can be found in McNab et all, 2009. The total capital cost was as follows:

  • Option 1 – $ 346.6M
  • Option 2 – $ 356.9M
  • Option 3 – $ 321.3M
  • Option 4 – $ 312.6M

The total estimated CAPEX for each circuit is within 14 percent, which infers that none of the options is a standout from a capital cost perspective at the accuracy level for this study. In comparison, the Boddington copper gold project CAPEX (Seidel et al. 2006) for the HPGR circuit option was 7 percent higher than the SAG option. Therefore, it appears that there may not be any significant CAPEX “penalty” for the adoption of more energy-efficient grinding technologies when considering magnetite ore processing.

Financial comparison

Figure 5. Project Net Present Value (NPV) comparisonHigh-level, pre-tax, net present value (NPV) determinations were calculated for Options 1 to 3 relative to the base case, Option 4 by applying a 10-percent discount rate over 12 years of operation. Option 4 was used as the base case since it returned the lowest capital and operating cost, and therefore NPV. Options 1 and 3 have a similar NPV outcome ranging between negative $94-95 M relative to Option 4. Option 2 shows the least favorable outcome with a $118 M NPV deficit relative to Option 4. This option has the combined disadvantages of both high capital and operating costs. The conclusion drawn from this financial evaluation is that highly energy-efficient autogenous processing routes can offer significant financial advantages for competent magnetite ores requiring fine grinding.


In this study it was found that highly energy-efficient autogenous processing routes can offer significant benefits for fine-grained, competent magnetite ores.

The traditional AG mill and pebble mill-style comminution circuit or those requiring significant steel grinding media to operate have been found to be less effective from a purely economic perspective. Circuit options utilizing multi-stage magnetic separation and with energy-efficient autogenous comminution equipment, although more complex, are more likely to add project value. For the ore type evaluated, the application of HPGR and stirred mill technology is indicated to reduce energy consumption by up to 25 percent compared with conventional flowsheets with wet tumbling mills.

There are many other flowsheet selection drivers that can become relevant, however, the operating cost associated with power draw and grinding media will always remain critical, even more so with the expected introduction of a carbon tax. A “synergy” of HPGR, pebble and stirred milling can result in a very effective circuit from a capital and operating point of view. It can be expected that highly energy-efficient autogenous processing routes would be further developed and increasingly applied in practice.

Related products