Search

The flotation circuit of Teck's new copper concentrator at Carmen de Andacollo, Chile, consists of the following stages: pre-rougher, pre-rougher cleaner, rougher, first cleaner, second cleaner and scavenger, with mechanical cells of 10, 20, 40, 100 and 200 m3, operating with forced air. The plant also includes a regrinding stage in a vertical mill of 1 500 HP. Measurement results along with analysis of data sampling are presented. The hydrodynamic and metallurgical behaviour of the flotation plant as well as the effects of different flotation stages on the overall plant performance were evaluated. For this purpose, three overall samplings and mass balances were developed and flotation kinetics identified for the overall solids as well as for particle size classes. Residence time distribution, bubble size distribution, local gas rate and froth depth were also measured. The plant instrumentation was evaluated to validate flowrate and froth depth measurements along the circuits. The mean treatment capacity during sampling was near 90% of the design specification due to mill constraints. The feed grade was 0.54-0.65% Cu, concentrate feed grades 26.5-29.2% Cu, while the plant copper recovery was 91.2-94.1% of Cu. From this evaluation, opportunities for improvement in operation and process control were identified., The flotation circuit of Teck's new copper concentrator at Carmen de Andacollo, Chile, consists of the following stages: pre-rougher, pre-rougher cleaner, rougher, first cleaner, second cleaner and scavenger, with mechanical cells of 10, 20, 40, 100 and 200 m3, operating with forced air. The plant also includes a regrinding stage in a vertical mill of 1 500 HP. Measurement results along with analysis of data sampling are presented. The hydrodynamic and metallurgical behaviour of the flotation plant as well as the effects of different flotation stages on the overall plant performance were evaluated. For this purpose, three overall samplings and mass balances were developed and flotation kinetics identified for the overall solids as well as for particle size classes. Residence time distribution, bubble size distribution, local gas rate and froth depth were also measured. The plant instrumentation was evaluated to validate flowrate and froth depth measurements along the circuits. The mean treatment capacity during sampling was near 90% of the design specification due to mill constraints. The feed grade was 0.54-0.65% Cu, concentrate feed grades 26.5-29.2% Cu, while the plant copper recovery was 91.2-94.1% of Cu. From this evaluation, opportunities for improvement in operation and process control were identified.

Laboratory flotation tests are completed by metallurgists to investigate the impact of changing one parameter from the standard condition to determine its impact on the metallurgical performance. These laboratory tests should be designed and executed so that they yield an indication of the potential benefit to the plant. If due care is taken in designing and executing the test as well as using the correct analysis techniques, then the conclusions are valid as they give the metallurgist an estimate of the potentially improved concentrate grades and recoveries, allowing the potential improvement in revenue to be estimated. Further, if these data are applied to a scale-up technique it is possible to provide a prediction of how the plant might behave if the test conditions were adopted within the operation. There is more to analysing the resultant flotation test data than comparing the final recovery point; this is a superficial analysis and tells the investigator very little about the grade/recovery/selectivity relationships or the flotation kinetics. It is possible to produce a quantitative comparison between a standard and a test condition; when these analysis techniques are combined with statistical methods and scale-up methods the data become very powerful., Laboratory flotation tests are completed by metallurgists to investigate the impact of changing one parameter from the standard condition to determine its impact on the metallurgical performance. These laboratory tests should be designed and executed so that they yield an indication of the potential benefit to the plant. If due care is taken in designing and executing the test as well as using the correct analysis techniques, then the conclusions are valid as they give the metallurgist an estimate of the potentially improved concentrate grades and recoveries, allowing the potential improvement in revenue to be estimated. Further, if these data are applied to a scale-up technique it is possible to provide a prediction of how the plant might behave if the test conditions were adopted within the operation. There is more to analysing the resultant flotation test data than comparing the final recovery point; this is a superficial analysis and tells the investigator very little about the grade/recovery/selectivity relationships or the flotation kinetics. It is possible to produce a quantitative comparison between a standard and a test condition; when these analysis techniques are combined with statistical methods and scale-up methods the data become very powerful.

The modern scale-up systems for ball mills require a series of grinding tests to determine population balance parameters that can be used for predicting circuit capacity for any ball mill size, feed rate and product specification. The time required to run the batch grinding tests is similar to that of a Bond ball mill Work Index determination. However, the calculations involved in determining the population balance parameters are considerably more elaborate, producing a number of selection and breakage function parameters for each sample. Typical variability campaigns carried out by Vale in green-field projects include hundreds of samples from drill cores. In order to assess the orebody variability with respect to ball mill grinding from population balance parameters hundreds of simulations would have to be carried out from independently determined sets of parameters, each one producing specific energy consumption under strictly specified conditions. This would be too elaborate and time-demanding and an alternative is simply to determine the Bond Work Index (WI) of each drill-core sample. The distribution of Bond WIs can then be used as a grindability parameter, and the population balance scale-up procedure needs to be carried out only for the sample that has the so-called representative WI. This procedure is demonstrated in this paper., The modern scale-up systems for ball mills require a series of grinding tests to determine population balance parameters that can be used for predicting circuit capacity for any ball mill size, feed rate and product specification. The time required to run the batch grinding tests is similar to that of a Bond ball mill Work Index determination. However, the calculations involved in determining the population balance parameters are considerably more elaborate, producing a number of selection and breakage function parameters for each sample. Typical variability campaigns carried out by Vale in green-field projects include hundreds of samples from drill cores. In order to assess the orebody variability with respect to ball mill grinding from population balance parameters hundreds of simulations would have to be carried out from independently determined sets of parameters, each one producing specific energy consumption under strictly specified conditions. This would be too elaborate and time-demanding and an alternative is simply to determine the Bond Work Index (WI) of each drill-core sample. The distribution of Bond WIs can then be used as a grindability parameter, and the population balance scale-up procedure needs to be carried out only for the sample that has the so-called representative WI. This procedure is demonstrated in this paper.

A dynamic simulator can show responses to operator actions such as setpoint changes, manual controls and mode changes. Dynamic simulation can reveal the target plant performance against process disturbances initiated by operator actions, interacting control loops, actuator and sensor malfunctions and quality changes of incoming feed material. It outperforms stationary static simulation by revealing transient responses between initial and final steadystates. A crushing plant has feeders, feed chutes, hoppers, crushers, screens and conveyors for processing aggregates. It is operated using information from flow and level sensors in addition to crusher-specific measurement information such as crusher setting, speed, power and pressure. Consequently, material flows and levels are the main variables that need to be modelled for dynamic simulation. They can be mainly treated using first principle models with empirical assumptions. To provide reasonable simulation accuracy for screening and crushing with bypass and recycling flows, size distribution behaviour is modelled using static models but flows themselves are treated using dynamic models. This paper shows how a dynamic flow and level simulator for a mobile crushing plant can be built up for process control design and concept development. Simulation results of a real mobile three-stage crushing plant are given., A dynamic simulator can show responses to operator actions such as setpoint changes, manual controls and mode changes. Dynamic simulation can reveal the target plant performance against process disturbances initiated by operator actions, interacting control loops, actuator and sensor malfunctions and quality changes of incoming feed material. It outperforms stationary static simulation by revealing transient responses between initial and final steadystates. A crushing plant has feeders, feed chutes, hoppers, crushers, screens and conveyors for processing aggregates. It is operated using information from flow and level sensors in addition to crusher-specific measurement information such as crusher setting, speed, power and pressure. Consequently, material flows and levels are the main variables that need to be modelled for dynamic simulation. They can be mainly treated using first principle models with empirical assumptions. To provide reasonable simulation accuracy for screening and crushing with bypass and recycling flows, size distribution behaviour is modelled using static models but flows themselves are treated using dynamic models. This paper shows how a dynamic flow and level simulator for a mobile crushing plant can be built up for process control design and concept development. Simulation results of a real mobile three-stage crushing plant are given.

Dry high-pressure grinding roll (HPGR) processing is gathering interest from industries that are operating in areas where water scarcity or process conditions require a minimising of water addition and energy consumption. HPGR technology could provide solutions even in operations where a very fine grind is required, by applying closed circuit arrangements with air classification, especially in the case where an integrated arrangement is used such as with the COMFLEX plant. Building on available processes and experiences from cement processing, tailored products can be generated at high throughput rates. This publication summarises some of the features and experiences with dry HPGR processing. Descriptive iron ore case studies are presented illustrating the process steps and grinding equipment arrangement to provide for fine products in a range of 20-125 micrometers from a feed size of 50 mm., Dry high-pressure grinding roll (HPGR) processing is gathering interest from industries that are operating in areas where water scarcity or process conditions require a minimising of water addition and energy consumption. HPGR technology could provide solutions even in operations where a very fine grind is required, by applying closed circuit arrangements with air classification, especially in the case where an integrated arrangement is used such as with the COMFLEX plant. Building on available processes and experiences from cement processing, tailored products can be generated at high throughput rates. This publication summarises some of the features and experiences with dry HPGR processing. Descriptive iron ore case studies are presented illustrating the process steps and grinding equipment arrangement to provide for fine products in a range of 20-125 micrometers from a feed size of 50 mm.

Innovative grinding media with a shape resembling a spheroidal Reuleaux tetrahedron or Relo body, are technologically suitable for production through casting, hot die forging and rolling. In contrast to the spherical ball with only one spherical surface, the Relo body ball has 4 spherical surfaces, 4 vertices and 6 edges. As a result, the productivity of Relo body grinding media is over 8% greater than that of spherical balls. The geometric, kinematic and production parameters of those bodies are investigated. It has been shown that the spherical surface radius of a Relo body ball is 2.14 times greater than the radius of an identical spherical ball; the 4 vertices and 6 edges of the Relo body creates conditions for more effective milling by crushing bigger particles of the material than spherical bodies. The shape of the Relo body allows greater volumes of hardened mass to be obtained with the use of low alloy steel and cheap technologies for heat treatment; the same technologies and materials are used for the production of the Relo body as for the spherical grinding bodies, so the price would be similar. Considering the great amount of energy spent for milling worldwide, the increase of productivity using Relo bodies instead of spherical bodies could have a significant economic and ecological impact., Innovative grinding media with a shape resembling a spheroidal Reuleaux tetrahedron or Relo body, are technologically suitable for production through casting, hot die forging and rolling. In contrast to the spherical ball with only one spherical surface, the Relo body ball has 4 spherical surfaces, 4 vertices and 6 edges. As a result, the productivity of Relo body grinding media is over 8% greater than that of spherical balls. The geometric, kinematic and production parameters of those bodies are investigated. It has been shown that the spherical surface radius of a Relo body ball is 2.14 times greater than the radius of an identical spherical ball; the 4 vertices and 6 edges of the Relo body creates conditions for more effective milling by crushing bigger particles of the material than spherical bodies. The shape of the Relo body allows greater volumes of hardened mass to be obtained with the use of low alloy steel and cheap technologies for heat treatment; the same technologies and materials are used for the production of the Relo body as for the spherical grinding bodies, so the price would be similar. Considering the great amount of energy spent for milling worldwide, the increase of productivity using Relo bodies instead of spherical bodies could have a significant economic and ecological impact.

The task of sizing a thickener for a defined application can result in a laborious job if the desired results are not clearly defined and the equipment utilised for this sizing is not adequate. Many parameters impact and affect the results when sizing a thickener, so a well structured test-work programme and a closer industrial simulation will provide a more accurate design. Sizing methods based on theoretical formulae or benchmarking are quite inaccurate since the safety factors applied to the obtained results increase the size of the thickeners, resulting in solutions with high capital and operating costs. Within the last five years and clearly also in the coming ones many high-throughput projects have been and will be developed in Chilea; all of them will have a need for lower capital expenditure, lower fresh water consumption, higher water recovery and lower operational expenses, but most of the time all these parameters are directly related to thickeners that are being designed from the theoretical point of view as expensive and large-diameter. The correct dynamic testing has shown that lower-diameter thickeners are perfectly adequate to work with the same throughput in an efficient manner and result in higher water recovery with lower capital and operating costs., The task of sizing a thickener for a defined application can result in a laborious job if the desired results are not clearly defined and the equipment utilised for this sizing is not adequate. Many parameters impact and affect the results when sizing a thickener, so a well structured test-work programme and a closer industrial simulation will provide a more accurate design. Sizing methods based on theoretical formulae or benchmarking are quite inaccurate since the safety factors applied to the obtained results increase the size of the thickeners, resulting in solutions with high capital and operating costs. Within the last five years and clearly also in the coming ones many high-throughput projects have been and will be developed in Chilea; all of them will have a need for lower capital expenditure, lower fresh water consumption, higher water recovery and lower operational expenses, but most of the time all these parameters are directly related to thickeners that are being designed from the theoretical point of view as expensive and large-diameter. The correct dynamic testing has shown that lower-diameter thickeners are perfectly adequate to work with the same throughput in an efficient manner and result in higher water recovery with lower capital and operating costs.

Significant in-process inventories can be locked in stockpiles of ore, concentrates, slags or reverts, or in process equipment such as bins, holding tanks, thickeners, heap leach pads and trucks, in and around mineral processing plants. Unfortunately, available measurements for such inventories in industrial practice are typically few and often carry significant inaccuracies. Ignoring estimation errors often leads to incoherencies in the plant production metal balances. As a consequence, in-process inventory levels are often adjusted, significantly and subjectively, in order to report coherent plant production metal balances. Operating plant production figures often show that such adjustments can have a significant financial impact. Following a brief review of the characteristics of common in-process inventories, an analysis of accounting methods based on industrial practice is presented along with a literature review. Although different methods can be applied, discussion focuses on the indirect inventory estimation method, applied when direct measurements of inventory are technically and economically challenging and/or inaccurate. The indirect estimation method uses feed and reclaim transactions to estimate in-process inventories. Its application, however, comes with significant operational and financial risks that should not be ignored., Significant in-process inventories can be locked in stockpiles of ore, concentrates, slags or reverts, or in process equipment such as bins, holding tanks, thickeners, heap leach pads and trucks, in and around mineral processing plants. Unfortunately, available measurements for such inventories in industrial practice are typically few and often carry significant inaccuracies. Ignoring estimation errors often leads to incoherencies in the plant production metal balances. As a consequence, in-process inventory levels are often adjusted, significantly and subjectively, in order to report coherent plant production metal balances. Operating plant production figures often show that such adjustments can have a significant financial impact. Following a brief review of the characteristics of common in-process inventories, an analysis of accounting methods based on industrial practice is presented along with a literature review. Although different methods can be applied, discussion focuses on the indirect inventory estimation method, applied when direct measurements of inventory are technically and economically challenging and/or inaccurate. The indirect estimation method uses feed and reclaim transactions to estimate in-process inventories. Its application, however, comes with significant operational and financial risks that should not be ignored.

The target of this investigation was to evaluate the application potential of anionic collectors obtained from vegetable oils in direct flotation of iron ore. The composition in fatty acids of four vegetable species (buriti, passion fruit, inajá and brazil nut) abundant in the Amazon Forest was determined via gas chromatography and correlated with their performance in flotation. Prior to their use, the oils were hydrolysed with sodium hydroxide and the reaction extent was assessed via Fourier transform infrared spectroscopy. The critical micelle concentration, calculated from surface tension determinations, and zeta potential measurements, using the microelectrophoretic technique, were used as tools to identify the adsorption mechanism of each collector on the haematite surfaces. Microflotation experiments were performed in a modified Hallimond tube using high-purity haematite and quartz samples. The highest haematite floatability levels were achieved in the pH range from 6.5 to 8.5. At the concentration of 2.5 mg/l, all collectors provided good selectivity between haematite and quartz. The results indicated the possibility of applying these Amazon vegetable species in the direct flotation of iron ores as full or partial replacements of the currently used collectors., The target of this investigation was to evaluate the application potential of anionic collectors obtained from vegetable oils in direct flotation of iron ore. The composition in fatty acids of four vegetable species (buriti, passion fruit, inajá and brazil nut) abundant in the Amazon Forest was determined via gas chromatography and correlated with their performance in flotation. Prior to their use, the oils were hydrolysed with sodium hydroxide and the reaction extent was assessed via Fourier transform infrared spectroscopy. The critical micelle concentration, calculated from surface tension determinations, and zeta potential measurements, using the microelectrophoretic technique, were used as tools to identify the adsorption mechanism of each collector on the haematite surfaces. Microflotation experiments were performed in a modified Hallimond tube using high-purity haematite and quartz samples. The highest haematite floatability levels were achieved in the pH range from 6.5 to 8.5. At the concentration of 2.5 mg/l, all collectors provided good selectivity between haematite and quartz. The results indicated the possibility of applying these Amazon vegetable species in the direct flotation of iron ores as full or partial replacements of the currently used collectors.

An alternative beneficiation process to recover nickel minerals from ores grading around 0.3% Ni was studied. The sample is composed basically of pyrrhotite and pentlandite with a fine degree of liberation. Magnetite and silicates, mostly talc and amphibolites, are the main gangue minerals. Based on the sample mineralogy, concentration operations such as magnetic separation and tabling were tested to remove part of the gangue minerals, increasing Ni concentration and promoting a reduction in the amount of sample to be processed by flotation. Particle size distribution was controlled during grinding to avoid production of an excess of fine particles by overgrinding. Slimes smaller than 20 mm were removed by hydrocyclones and the finely ground material (-105 +20 mm) was processed as 17-20% solid feed by column flotation in a 20 ft high pilot column. The product obtained in the rougher step reached an average grade of 1.5% Ni, with 50% global recovery and 90% recovery in the flotation operation. These are interesting results given that there was only one flotation stage. The goal of the process was to study the feasibility of using low-grade nickel ores for the production of concentrates grading up to 2.5% Ni to be used as feed in the metallurgical production of nickel matte., An alternative beneficiation process to recover nickel minerals from ores grading around 0.3% Ni was studied. The sample is composed basically of pyrrhotite and pentlandite with a fine degree of liberation. Magnetite and silicates, mostly talc and amphibolites, are the main gangue minerals. Based on the sample mineralogy, concentration operations such as magnetic separation and tabling were tested to remove part of the gangue minerals, increasing Ni concentration and promoting a reduction in the amount of sample to be processed by flotation. Particle size distribution was controlled during grinding to avoid production of an excess of fine particles by overgrinding. Slimes smaller than 20 mm were removed by hydrocyclones and the finely ground material (-105 +20 mm) was processed as 17-20% solid feed by column flotation in a 20 ft high pilot column. The product obtained in the rougher step reached an average grade of 1.5% Ni, with 50% global recovery and 90% recovery in the flotation operation. These are interesting results given that there was only one flotation stage. The goal of the process was to study the feasibility of using low-grade nickel ores for the production of concentrates grading up to 2.5% Ni to be used as feed in the metallurgical production of nickel matte.

This paper presents a conceptual evaluation of silver recovery with tailings polymetallic flotation. Minera San Cristobal ore from Bolivia contains significant amounts of Pb, Zn and Ag sulphide, resulting in good recoveries by flotation. Options are being considered to optimise the silver recovery from tailings flotation which would bring significant economic benefit because of the high Ag price and the large capacity of the plant. Experimental tests were performed with polymetallic flotation tailings assaying between 15 and 25 g/t Ag. Concentration by flotation was found to be the preferred primary recovery route. The concentrate assayed between 100 and 200 g/t Ag in rougher flotation, with 68% recovery for particles smaller than 75 micrometres. Pretreatment of the flotation tailings indicated high Ag recoveries could be achieved by flotation, producing a high-grade concentrate. Preliminary terms for the sale of flotation concentrate were discussed and processing of the flotation concentrate was considered to be more economical, as treatment on site had the advantages of requiring a small leaching and recovery facility. The treatment of flotation concentrate was performed by cyanide leaching rather than acid chloride leaching. Using the Factor Split tool, grades and recoveries at industrial level were predicted. Results have shown construction of a 16 000 tpd flotation/leaching plant requires a total investment of US, This paper presents a conceptual evaluation of silver recovery with tailings polymetallic flotation. Minera San Cristobal ore from Bolivia contains significant amounts of Pb, Zn and Ag sulphide, resulting in good recoveries by flotation. Options are being considered to optimise the silver recovery from tailings flotation which would bring significant economic benefit because of the high Ag price and the large capacity of the plant. Experimental tests were performed with polymetallic flotation tailings assaying between 15 and 25 g/t Ag. Concentration by flotation was found to be the preferred primary recovery route. The concentrate assayed between 100 and 200 g/t Ag in rougher flotation, with 68% recovery for particles smaller than 75 micrometres. Pretreatment of the flotation tailings indicated high Ag recoveries could be achieved by flotation, producing a high-grade concentrate. Preliminary terms for the sale of flotation concentrate were discussed and processing of the flotation concentrate was considered to be more economical, as treatment on site had the advantages of requiring a small leaching and recovery facility. The treatment of flotation concentrate was performed by cyanide leaching rather than acid chloride leaching. Using the Factor Split tool, grades and recoveries at industrial level were predicted. Results have shown construction of a 16 000 tpd flotation/leaching plant requires a total investment of US

Minera Fresnillo and Minera Saucito process minerals with gold, silver, lead and zinc by means of selective flotation, to obtain lead concentrates with most of the gold-silver and zinc concentrates having relatively high values of these elements. Gold and silver recovery in both concentrates is around 85% and 91% respectively and the remaining values show as inclusions of pyrite and non-sulphide gangue less than 10 micrometres in size. With the purpose of maximising recoveries, tests were done to float pyrite using tailings of selective flotation, achieving concentrate grades around 3 g/ton Au and 700 g/t Ag, with 70% recovery for both elements. By cyanidation of the concentrate, without grinding it, 51% gold and 60% silver were recovered and with regrinding to 8 microns recoveries were reached of 73% and 91% respectively. Based on these results, it was decided to float pyrite from the lead-zinc tailings and to build a plant to leach this concentrate, so as to recover around 8% of additional Au and 6% of additional Ag from the total gold and silver in the main plant feed., Minera Fresnillo and Minera Saucito process minerals with gold, silver, lead and zinc by means of selective flotation, to obtain lead concentrates with most of the gold-silver and zinc concentrates having relatively high values of these elements. Gold and silver recovery in both concentrates is around 85% and 91% respectively and the remaining values show as inclusions of pyrite and non-sulphide gangue less than 10 micrometres in size. With the purpose of maximising recoveries, tests were done to float pyrite using tailings of selective flotation, achieving concentrate grades around 3 g/ton Au and 700 g/t Ag, with 70% recovery for both elements. By cyanidation of the concentrate, without grinding it, 51% gold and 60% silver were recovered and with regrinding to 8 microns recoveries were reached of 73% and 91% respectively. Based on these results, it was decided to float pyrite from the lead-zinc tailings and to build a plant to leach this concentrate, so as to recover around 8% of additional Au and 6% of additional Ag from the total gold and silver in the main plant feed.

Over the past decade the growth and development of mining projects has increased dramatically, as have the requirements to design and develop grinding circuits, mainly in the field of SAG and ball mill circuits. There are several design methodologies available on the market or through the use of commercial software or expert consultants. This article presents a comparative analysis of three methods for grinding circuit design widely used in the market today, Millpower 2000, JKSimMet and Aminpro, applied to a new mining sector of a current Anglo American mine. The parameters were obtained from 100 drill core composites, each representing 30 m of drill core, in a sector that represents the first five years of production for a new concentrator. An important feature of this material is its high hardness for SAG grinding (140 min. average SAG power index or SPI - time to grind from 80% -12.5 mm to 80% -1.70 mm), and this perhaps provides an additional advantage to the study as it represents an extreme condition. The analysis compares the resulting equipment sizes for each of the three methodologies, considering a proposed tonnage of 90 000 t/d at 80% passing 180 micrometres. (Authors.), Over the past decade the growth and development of mining projects has increased dramatically, as have the requirements to design and develop grinding circuits, mainly in the field of SAG and ball mill circuits. There are several design methodologies available on the market or through the use of commercial software or expert consultants. This article presents a comparative analysis of three methods for grinding circuit design widely used in the market today, Millpower 2000, JKSimMet and Aminpro, applied to a new mining sector of a current Anglo American mine. The parameters were obtained from 100 drill core composites, each representing 30 m of drill core, in a sector that represents the first five years of production for a new concentrator. An important feature of this material is its high hardness for SAG grinding (140 min. average SAG power index or SPI - time to grind from 80% -12.5 mm to 80% -1.70 mm), and this perhaps provides an additional advantage to the study as it represents an extreme condition. The analysis compares the resulting equipment sizes for each of the three methodologies, considering a proposed tonnage of 90 000 t/d at 80% passing 180 micrometres. (Authors.)

Hydrocyclones are devices used worldwide in mineral processing for desliming, classification, selective classification, thickening and pre-concentration. Versatile in application, the hydrocyclone is the standard classifier in closed circuit milling. A hydrocyclone is composed of a cylindrical and a conical section joined together, without any moving parts and is capable of performing granular material separation in pulp. The pulp is fed under pressure tangentially to the cylindrical section. The granular separation mechanism is complex and its mathematical modelling empirical. The most used model for hydrocyclone dimensioning was proposed by Plitt in 1976. Combining the first industrial database on cyclones generated at the Julius Kruttschnitt Mineral Research Centre with his own laboratory data, Plitt developed an alternative general-purpose cyclone model. The present paper shows a modification in the Plitt´s model constant, obtained by exponential regression of simulated data for three different hydrocyclone geometries, Rietema, Bradley and Krebs. The proposed model validation used literature data on phosphate ore with fifteen different hydrocyclone geometries and shows a correlation equal to 88.2% between experimental and calculated corrected cut size, whereas the correlation obtained using Plitt’s unmodified model was only 11.5%., Hydrocyclones are devices used worldwide in mineral processing for desliming, classification, selective classification, thickening and pre-concentration. Versatile in application, the hydrocyclone is the standard classifier in closed circuit milling. A hydrocyclone is composed of a cylindrical and a conical section joined together, without any moving parts and is capable of performing granular material separation in pulp. The pulp is fed under pressure tangentially to the cylindrical section. The granular separation mechanism is complex and its mathematical modelling empirical. The most used model for hydrocyclone dimensioning was proposed by Plitt in 1976. Combining the first industrial database on cyclones generated at the Julius Kruttschnitt Mineral Research Centre with his own laboratory data, Plitt developed an alternative general-purpose cyclone model. The present paper shows a modification in the Plitt´s model constant, obtained by exponential regression of simulated data for three different hydrocyclone geometries, Rietema, Bradley and Krebs. The proposed model validation used literature data on phosphate ore with fifteen different hydrocyclone geometries and shows a correlation equal to 88.2% between experimental and calculated corrected cut size, whereas the correlation obtained using Plitt’s unmodified model was only 11.5%.

The selection of solid/liquid separation is often made by treating the unit operations of sedimentation, thickening and filtration separately, defining a nominal performance then minimising each item's capital expenditure. The degree to which the thickener performance affects the filtration performance and how filtrates and wash waters are returned to the thickeners can have a significant effect on the performance of both technologies. Further complicating the selection are the storage and pumping of thickened slurries, which modify the performance and sizing of the dewatering technologies. Solid/liquid separation performance, like metallurgy, is defined by the process conditions and the optimum conditions are often vastly different. In the never-ending pursuit of reducing costs, the logical point to study is the larger-cost areas. Where solid/liquid separation costs are high, optimising the process conditions and redefining the unit operation duties may provide benefits. To help separation in most proposed flowsheets it is worth looking at the logic behind the selection of equipment. This paper reviews existing solutions in case studies and alternative theoretical selections., The selection of solid/liquid separation is often made by treating the unit operations of sedimentation, thickening and filtration separately, defining a nominal performance then minimising each item's capital expenditure. The degree to which the thickener performance affects the filtration performance and how filtrates and wash waters are returned to the thickeners can have a significant effect on the performance of both technologies. Further complicating the selection are the storage and pumping of thickened slurries, which modify the performance and sizing of the dewatering technologies. Solid/liquid separation performance, like metallurgy, is defined by the process conditions and the optimum conditions are often vastly different. In the never-ending pursuit of reducing costs, the logical point to study is the larger-cost areas. Where solid/liquid separation costs are high, optimising the process conditions and redefining the unit operation duties may provide benefits. To help separation in most proposed flowsheets it is worth looking at the logic behind the selection of equipment. This paper reviews existing solutions in case studies and alternative theoretical selections.

The objective of this research was to study the feasibility of the flotation process for the recovery of lead from a metallurgical slag composed of 16.9% PbO, 29.8% SO3 and 41.6% Fe2O3. Flotation tests were performed on a pilot scale with a processing capacity of 40 kg/h, using metallurgical slag samples from Acumuladores Moura, an automotive battery manufacturer. The target was to obtain a final concentrate grading up to 60% Pb. Several gravity separation methods were conducted to set up and evaluate process routes for a further recovery or concentrate enrichment. Initially a pilot flotation circuit was assembled with a capacity of 50 kg/h of slag (dry base). The initial reagent system was composed of an anionic collector (xanthate) and a frother. A sulphuric acid solution was used as pH conditioner. The gravity concentration tests were performed in a hydraulic classifier. The classification experiments were run in an Inbras elutriator, shaking table and spiral concentrators. In the pilot flotation tests it was possible to produce concentrates grading 50% Pb and metallurgical recoveries up to 80% of Pb. The flotation concentrate was reprocessed on the shaking table and a new product obtained with a grade up to 70% Pb, thereby reaching the initial goals of this work., The objective of this research was to study the feasibility of the flotation process for the recovery of lead from a metallurgical slag composed of 16.9% PbO, 29.8% SO3 and 41.6% Fe2O3. Flotation tests were performed on a pilot scale with a processing capacity of 40 kg/h, using metallurgical slag samples from Acumuladores Moura, an automotive battery manufacturer. The target was to obtain a final concentrate grading up to 60% Pb. Several gravity separation methods were conducted to set up and evaluate process routes for a further recovery or concentrate enrichment. Initially a pilot flotation circuit was assembled with a capacity of 50 kg/h of slag (dry base). The initial reagent system was composed of an anionic collector (xanthate) and a frother. A sulphuric acid solution was used as pH conditioner. The gravity concentration tests were performed in a hydraulic classifier. The classification experiments were run in an Inbras elutriator, shaking table and spiral concentrators. In the pilot flotation tests it was possible to produce concentrates grading 50% Pb and metallurgical recoveries up to 80% of Pb. The flotation concentrate was reprocessed on the shaking table and a new product obtained with a grade up to 70% Pb, thereby reaching the initial goals of this work.

When deciding between flotation flowsheets with similar capital outlay, the key performance indicators are usually the predicted grade and recovery of the primary metal for beneficiation. However, some flowsheets are more sensitive to changes in ore grade and kinetics than others. Since ore characteristics cannot be perfectly known, sensitivity to ore variability presents risk to the project viability. A second source of risk comes from inadequately accounting for disturbances to circuit operation. This work presents a case study where two flotation flowsheets were compared during the final flowsheet design phase for a large Chilean copper concentrator. The flowsheets were calibrated from rougher flotation tests, cleaner tests and historical plant benchmarks and presented similar flotation performance in terms of Cu recovery and grade at the baseline conditions. The Integrated Geometallurgical Simulator was used to conduct a full Monte Carlo sensitivity analysis by varying the ore characteristics and the operation of the circuits in the flotation simulations. The results of the analysis showed a clear preference between the two flowsheets, one predicting consistent results whereas the other showed a significant downside risk from the predicted aggregate performance., When deciding between flotation flowsheets with similar capital outlay, the key performance indicators are usually the predicted grade and recovery of the primary metal for beneficiation. However, some flowsheets are more sensitive to changes in ore grade and kinetics than others. Since ore characteristics cannot be perfectly known, sensitivity to ore variability presents risk to the project viability. A second source of risk comes from inadequately accounting for disturbances to circuit operation. This work presents a case study where two flotation flowsheets were compared during the final flowsheet design phase for a large Chilean copper concentrator. The flowsheets were calibrated from rougher flotation tests, cleaner tests and historical plant benchmarks and presented similar flotation performance in terms of Cu recovery and grade at the baseline conditions. The Integrated Geometallurgical Simulator was used to conduct a full Monte Carlo sensitivity analysis by varying the ore characteristics and the operation of the circuits in the flotation simulations. The results of the analysis showed a clear preference between the two flowsheets, one predicting consistent results whereas the other showed a significant downside risk from the predicted aggregate performance.

Process water used in mineral processing plants is usually a blend of different water supplies, including fresh water, bore and surface waters, waters from treated or untreated sewage plants and industrial effluents, and recycled waters. The quality and more particularly their constituents may impact flotation performance. The objective of this study is to investigate the effect of various natural acidic organic polymers, such as humic and fulvic acids, and tannic acid on the flotation of Cu and Mo sulphides in alkaline conditions (pH 9.3). Adsorption studies of these polymers on chalcopyrite and molybdenite were also conducted to interpret the flotation results. It was shown that Cu and Mo grades decreased with addition of humic, fulvic and tannic acid in that order, and the effect was more pronounced for molybdenite grade and recovery. UV-visible spectroscopy solution analysis also confirmed that these polymers were indeed adsorbed on the surface of molybdenite and chalcopyrite. The detrimental effect of organic materials in flotation is explained in terms of changes in froth characteristics, enhanced entrainment of fine particles, and a decrease in the overall hydrophobicity of the sulphide mineral surfaces, due to the adsorption of hydrophilic groups in these polymers which then prevent bubble-particle adhesion., Process water used in mineral processing plants is usually a blend of different water supplies, including fresh water, bore and surface waters, waters from treated or untreated sewage plants and industrial effluents, and recycled waters. The quality and more particularly their constituents may impact flotation performance. The objective of this study is to investigate the effect of various natural acidic organic polymers, such as humic and fulvic acids, and tannic acid on the flotation of Cu and Mo sulphides in alkaline conditions (pH 9.3). Adsorption studies of these polymers on chalcopyrite and molybdenite were also conducted to interpret the flotation results. It was shown that Cu and Mo grades decreased with addition of humic, fulvic and tannic acid in that order, and the effect was more pronounced for molybdenite grade and recovery. UV-visible spectroscopy solution analysis also confirmed that these polymers were indeed adsorbed on the surface of molybdenite and chalcopyrite. The detrimental effect of organic materials in flotation is explained in terms of changes in froth characteristics, enhanced entrainment of fine particles, and a decrease in the overall hydrophobicity of the sulphide mineral surfaces, due to the adsorption of hydrophilic groups in these polymers which then prevent bubble-particle adhesion.

Flotation of mineral particles represents a key process in the mining industry. This study aims at evaluating mathematical models which are able to describe the counter-current columnar flotation process. Thereby, from a modelling point of view two main approaches, analytical considerations and numerical modelling, are applied. Firstly, the flotation process is pictured by an analytical, one-dimensional drift-flux model. Starting with pure fluid dynamical considerations, it is possible to obtain the limit of homogeneous operation of columnar flotation. Next, based on a rigorous literature review, particle collision, attachment and detachment sub-models have been integrated. As a result, the efficiency of columnar flotation can be estimated as a function of the main operating conditions such as gas flow rate, mean bubble or particle diameter. In a second step the limits of homogeneous operation are investigated by means of numerical simulation. The computational fluid dynamics (CFD) tool OpenFOAM is used to examine the stability of the regime in homogeneously aerated bubble columns. The results of the CFD analysis are compared with the estimations for regime stability from the analytical approach. (Authors.), Flotation of mineral particles represents a key process in the mining industry. This study aims at evaluating mathematical models which are able to describe the counter-current columnar flotation process. Thereby, from a modelling point of view two main approaches, analytical considerations and numerical modelling, are applied. Firstly, the flotation process is pictured by an analytical, one-dimensional drift-flux model. Starting with pure fluid dynamical considerations, it is possible to obtain the limit of homogeneous operation of columnar flotation. Next, based on a rigorous literature review, particle collision, attachment and detachment sub-models have been integrated. As a result, the efficiency of columnar flotation can be estimated as a function of the main operating conditions such as gas flow rate, mean bubble or particle diameter. In a second step the limits of homogeneous operation are investigated by means of numerical simulation. The computational fluid dynamics (CFD) tool OpenFOAM is used to examine the stability of the regime in homogeneously aerated bubble columns. The results of the CFD analysis are compared with the estimations for regime stability from the analytical approach. (Authors.)

In general, efforts made to achieve optimisation and control of the real-time crushing process are rare. None of the currently existing models are truly useful for control system design, as they all limit process examination to the steady-state operation of the crusher even though a crushing process without a feedback control system is never in a steady state. In order to achieve the desired performance, control is needed as crushing plant dynamics are connected in complex ways and adjusting one device affects the entire plant, since the dynamics of different devices are directly connected. Sensors in a crushing environment are problematic, so information about the flow and properties of the rock material is difficult to obtain. Actuators for manipulating the process are lacking or nonlinear in nature, or otherwise unsuitable. The purpose of this paper is to present simple and implementable solutions that try to address the problems mentioned above. The paper also promotes the view that the process needs to be controlled as a whole, including all devices such as silos and conveyors, instead of focusing on optimising specific devices only; investing in special control solutions such as model-predictive control is not feasible unless the basic controls in a plant are running properly., In general, efforts made to achieve optimisation and control of the real-time crushing process are rare. None of the currently existing models are truly useful for control system design, as they all limit process examination to the steady-state operation of the crusher even though a crushing process without a feedback control system is never in a steady state. In order to achieve the desired performance, control is needed as crushing plant dynamics are connected in complex ways and adjusting one device affects the entire plant, since the dynamics of different devices are directly connected. Sensors in a crushing environment are problematic, so information about the flow and properties of the rock material is difficult to obtain. Actuators for manipulating the process are lacking or nonlinear in nature, or otherwise unsuitable. The purpose of this paper is to present simple and implementable solutions that try to address the problems mentioned above. The paper also promotes the view that the process needs to be controlled as a whole, including all devices such as silos and conveyors, instead of focusing on optimising specific devices only; investing in special control solutions such as model-predictive control is not feasible unless the basic controls in a plant are running properly.

High-pressure grinding roll (HPGR) technology is attracting much interest from the industry, with a multitude of projects being engaged and increasing numbers of units being applied in the minerals processing world. To enable the applicability of HPGR in mineral projects to be evaluated, but also to provide an opportunity for studying the HPGR process in detail, a KHD pilot unit (rolls of 0.8 m diameter x 0.25 m wide) and auxiliary equipment has been installed at the SGS-CIMM Tecnologías y Servicios Laboratory in Quilicura, Santiago, Chile. This facility has been in operation for test work and research and development for the South-American industry since December 2011. The paper summarises the basics of a general test-programmepproach and testing procedures with this pilot unit, and emphasises the necessity to focus also on materials handling properties of HPGR products such as flake desagglomeration and moisture content. A descriptive evaluation is presented illustrating the results of a HPGR variability test programme and the influence of ore density on an example ore tested at the laboratory in Quilicura., High-pressure grinding roll (HPGR) technology is attracting much interest from the industry, with a multitude of projects being engaged and increasing numbers of units being applied in the minerals processing world. To enable the applicability of HPGR in mineral projects to be evaluated, but also to provide an opportunity for studying the HPGR process in detail, a KHD pilot unit (rolls of 0.8 m diameter x 0.25 m wide) and auxiliary equipment has been installed at the SGS-CIMM Tecnologías y Servicios Laboratory in Quilicura, Santiago, Chile. This facility has been in operation for test work and research and development for the South-American industry since December 2011. The paper summarises the basics of a general test-programmepproach and testing procedures with this pilot unit, and emphasises the necessity to focus also on materials handling properties of HPGR products such as flake desagglomeration and moisture content. A descriptive evaluation is presented illustrating the results of a HPGR variability test programme and the influence of ore density on an example ore tested at the laboratory in Quilicura.

Spiral concentrators consist of a helical trough that spirals around a central column. The slurry enters at the top and minerals separate as a result of density differences, with more dense material concentrating in the central part and lighter material migrating towards the outer edge of the trough. As the number of turns in the spiral increases, the separation becomes more pronounced but significant amounts of fine heavies are lost to the tailings. This paper introduces initial research on observing the motion of particles in a laboratory spiral concentrator to enhance the understanding of spirals. The technique used is positron emission particle tracking, developed at the University of Birmingham to track single particles in engineering equipment. In order to track the position of a particle it first must be labelled with a radionuclide tracer, which decays through beta+ decay. As the radionuclide decays, positrons are emitted which annihilate with electrons producing two back-to-back gamma-rays, detected by a camera consisting of a pair of positron-sensitive sensors, and then a location algorithm is used to calculate the tracer location. Results of initial tests are showing the helical motion as tracking particles flow down, and the trajectories that the particles follow. This work will lead to the validation of models of spiral concentrators and optimised performance., Spiral concentrators consist of a helical trough that spirals around a central column. The slurry enters at the top and minerals separate as a result of density differences, with more dense material concentrating in the central part and lighter material migrating towards the outer edge of the trough. As the number of turns in the spiral increases, the separation becomes more pronounced but significant amounts of fine heavies are lost to the tailings. This paper introduces initial research on observing the motion of particles in a laboratory spiral concentrator to enhance the understanding of spirals. The technique used is positron emission particle tracking, developed at the University of Birmingham to track single particles in engineering equipment. In order to track the position of a particle it first must be labelled with a radionuclide tracer, which decays through beta+ decay. As the radionuclide decays, positrons are emitted which annihilate with electrons producing two back-to-back gamma-rays, detected by a camera consisting of a pair of positron-sensitive sensors, and then a location algorithm is used to calculate the tracer location. Results of initial tests are showing the helical motion as tracking particles flow down, and the trajectories that the particles follow. This work will lead to the validation of models of spiral concentrators and optimised performance.

Brazil has important occurrences of graphite linked to regional metamorphism caused by granitic intrusion during the Neoproterozoic Pan African-Brazilian orogenic cycle, commonly containing ore intergrowths between lamellae of graphite and kaolinite. Its processing practice is reviewed in order to contextualise a campaign for technological characterisation of some flake graphite ores from Guaratinga and Eunapolis regions, Bahia, Brazil. The orebodies are located inside the Aracuai orogenic belt. Jaw, roll and impact crushing, ball and rod grinding and flotation tests were carried out with representative samples from this area, both weathered rock and bore-hole probe. Flotation reagents successfully used were pine oil and diesel oil together at pH 8 and frother Flotanol D25 at the same pH value. The experiments have displayed promising results for future economic exploitation of orebodies recently discovered., Brazil has important occurrences of graphite linked to regional metamorphism caused by granitic intrusion during the Neoproterozoic Pan African-Brazilian orogenic cycle, commonly containing ore intergrowths between lamellae of graphite and kaolinite. Its processing practice is reviewed in order to contextualise a campaign for technological characterisation of some flake graphite ores from Guaratinga and Eunapolis regions, Bahia, Brazil. The orebodies are located inside the Aracuai orogenic belt. Jaw, roll and impact crushing, ball and rod grinding and flotation tests were carried out with representative samples from this area, both weathered rock and bore-hole probe. Flotation reagents successfully used were pine oil and diesel oil together at pH 8 and frother Flotanol D25 at the same pH value. The experiments have displayed promising results for future economic exploitation of orebodies recently discovered.

Effective dewatering of mineral waste tailings and subsequent disposal are important issues for the mining and mineral processing industry world-wide, as beneficiation, fine grinding, froth flotation and hydrometallurgical extraction from low-grade ores involve voluminous amounts of water, generating equivalently large volumes of low-solid-density mineral waste tailings. For clay and some oxide minerals, slimes dewatering, flocculation and thickening operations produce a fast settling rate but poorly consolidating sediments, even after several years of impoundment in dams, warranting alternative, cost-effective methods for enhanced dewaterability. In this work, electroosmotic dewatering and consolidation of model and real plant kaolinite and smectite clay slimes or pulps were investigated. The study focused on the specific influence of variables such as dispersion pH, supernatant (electrolyte) metal ion concentration, voltage input and current reversal/intermittence on energy consumption during electroosmosis. Optimum physicochemical conditions were established. A dramatic increase in sediment loading by 20-30 wt.% to spadeable pulp consistency was achieved at an acceptable power consumption rate., Effective dewatering of mineral waste tailings and subsequent disposal are important issues for the mining and mineral processing industry world-wide, as beneficiation, fine grinding, froth flotation and hydrometallurgical extraction from low-grade ores involve voluminous amounts of water, generating equivalently large volumes of low-solid-density mineral waste tailings. For clay and some oxide minerals, slimes dewatering, flocculation and thickening operations produce a fast settling rate but poorly consolidating sediments, even after several years of impoundment in dams, warranting alternative, cost-effective methods for enhanced dewaterability. In this work, electroosmotic dewatering and consolidation of model and real plant kaolinite and smectite clay slimes or pulps were investigated. The study focused on the specific influence of variables such as dispersion pH, supernatant (electrolyte) metal ion concentration, voltage input and current reversal/intermittence on energy consumption during electroosmosis. Optimum physicochemical conditions were established. A dramatic increase in sediment loading by 20-30 wt.% to spadeable pulp consistency was achieved at an acceptable power consumption rate.

Frequency and stroke amplitude are the main factors that have influences on screen dewatering efficiency, while knowledge of particulate systems characteristics is important to forecast the flowrate and the residual moisture. The study reports on an experimental campaign to determine moisture isovalues inside a dewatering screen bed, carried out on a laboratory scale. A mathematical model is proposed that demonstrates how the controlling mechanisms of the water drainage process actually work. The model enables the residual moisture’s evolution to be determined from retention time and bed porosity. Tests were carried out to validate the model, using glass beads and sand. A computer system was implemented to simulate many scenarios for dewatering screens, the results showing that the model describes the operations in an acceptable way but does not take into account the effects of capillarity and dynamic acceleration. New tests need to be carried out to consider the effects of capillaries, as well as using another screen that acts as if it were industrial., Frequency and stroke amplitude are the main factors that have influences on screen dewatering efficiency, while knowledge of particulate systems characteristics is important to forecast the flowrate and the residual moisture. The study reports on an experimental campaign to determine moisture isovalues inside a dewatering screen bed, carried out on a laboratory scale. A mathematical model is proposed that demonstrates how the controlling mechanisms of the water drainage process actually work. The model enables the residual moisture’s evolution to be determined from retention time and bed porosity. Tests were carried out to validate the model, using glass beads and sand. A computer system was implemented to simulate many scenarios for dewatering screens, the results showing that the model describes the operations in an acceptable way but does not take into account the effects of capillarity and dynamic acceleration. New tests need to be carried out to consider the effects of capillaries, as well as using another screen that acts as if it were industrial.

Mining companies, due to the gradual depletion of their reserves, have to grind their ore finer to achieve the same liberation degree, increasing the fine content of the concentrate. These fine particles are an important parameter during filtration and dewatering. Therefore, of additives to improve filtration and dewatering are increasingly used in mining operations in order to decrease time and material during the industrial process. The general mechanisms for dewatering and filtration are similar. Filtration occurs when the solid contained in the pulp is separated from the liquid through porous filter media and forms a cake. Dewatering is the process in which solid material containing liquid is conveyed or elevated in a manner that allows the liquid to drain off while the solid material is in transit. The water removal from the concentrate decreases the cost for the next handling step, for example, when it is processed at the mine location and needs to be transported. Dewatering and filtration laboratory tests were developed to evaluate different surfactants, and to find the best chemical composition. This was done through a quantitative treatment as well as investigating the way the water remains on the ore particle surface. Different ores were evaluated and different chemical solutions identified, probably due to the different mineralogy and fineness of the ores., Mining companies, due to the gradual depletion of their reserves, have to grind their ore finer to achieve the same liberation degree, increasing the fine content of the concentrate. These fine particles are an important parameter during filtration and dewatering. Therefore, of additives to improve filtration and dewatering are increasingly used in mining operations in order to decrease time and material during the industrial process. The general mechanisms for dewatering and filtration are similar. Filtration occurs when the solid contained in the pulp is separated from the liquid through porous filter media and forms a cake. Dewatering is the process in which solid material containing liquid is conveyed or elevated in a manner that allows the liquid to drain off while the solid material is in transit. The water removal from the concentrate decreases the cost for the next handling step, for example, when it is processed at the mine location and needs to be transported. Dewatering and filtration laboratory tests were developed to evaluate different surfactants, and to find the best chemical composition. This was done through a quantitative treatment as well as investigating the way the water remains on the ore particle surface. Different ores were evaluated and different chemical solutions identified, probably due to the different mineralogy and fineness of the ores.

Molybdenite (MoS2) is predominantly recovered from the flotation of copper porphyric ores, which is carried out at a high pH and achieved with lime, which supplies Ca ions to the flotation pulp. MoS2 is selectively floated from the Cu-MoS2 bulk concentrate while depressing the Cu minerals. In the Cu ore flotation circuit, the Cu recovery is high, but that of MoS2 is low. To delineate the loss of the MoS2 natural floatability and low recovery, the characteristics of the hydrophobic face of pristine MoS2 particles have been studied through atomic force microscopy, which revealed that the face is made up of microcrystals of MoS2 having well defined nano-edges and faces. The nano-edges render hydrophilic the faces of molybdenite particles and seem to be responsible for the low natural floatability of MoS2, the adsorption of Ca ions on MoS2 faces and the decrease of this natural floatability by Ca ions at a high pH. Both hydrophobic and hydrophilic surface fractions on the face have been determined using contact angle values determined by the captive bubble and Washburn capillary penetration methods. It was found that the hydrophobic surface fraction is only about 50%. The effect of pH and Ca ions on the natural floatability of various size fractions of MoS2 particles is presented in this study., Molybdenite (MoS2) is predominantly recovered from the flotation of copper porphyric ores, which is carried out at a high pH and achieved with lime, which supplies Ca ions to the flotation pulp. MoS2 is selectively floated from the Cu-MoS2 bulk concentrate while depressing the Cu minerals. In the Cu ore flotation circuit, the Cu recovery is high, but that of MoS2 is low. To delineate the loss of the MoS2 natural floatability and low recovery, the characteristics of the hydrophobic face of pristine MoS2 particles have been studied through atomic force microscopy, which revealed that the face is made up of microcrystals of MoS2 having well defined nano-edges and faces. The nano-edges render hydrophilic the faces of molybdenite particles and seem to be responsible for the low natural floatability of MoS2, the adsorption of Ca ions on MoS2 faces and the decrease of this natural floatability by Ca ions at a high pH. Both hydrophobic and hydrophilic surface fractions on the face have been determined using contact angle values determined by the captive bubble and Washburn capillary penetration methods. It was found that the hydrophobic surface fraction is only about 50%. The effect of pH and Ca ions on the natural floatability of various size fractions of MoS2 particles is presented in this study.

A new methodology for the design of flotation circuits is presented. It considers three design and analysis tools: process group contribution to generate a preliminary evaluation of flotation circuits; sensitivity analysis to identify critical variables and stages; and reverse simulation to guide experimentation and establishment of operating conditions. The methodology includes three steps: (i) definition and analysis of the problem: the feed material is characterised, the general criteria for circuit evaluation are defined and the value of recoveries in flotation stages and number of stages used are estimated; (ii) synthesis and screening of alternatives: all the feasible circuit alternatives are generated, the performance of each circuit is estimated using group contribution, circuits are classified based on criteria previously defined, and finally the performance of the most promising circuits is validated by modelling; (iii) improvements and final design: the equipment and operational conditions are determined using modelling and/or experimental tests, local sensitivity analysis is used to identify the most critical variables and reverse simulation is used for modelling. A case study illustrates the proposed methodology. This is a hybrid design method because empirical knowledge, approximate models and designer’s experience are used together with rigorous modelling and testing., A new methodology for the design of flotation circuits is presented. It considers three design and analysis tools: process group contribution to generate a preliminary evaluation of flotation circuits; sensitivity analysis to identify critical variables and stages; and reverse simulation to guide experimentation and establishment of operating conditions. The methodology includes three steps: (i) definition and analysis of the problem: the feed material is characterised, the general criteria for circuit evaluation are defined and the value of recoveries in flotation stages and number of stages used are estimated; (ii) synthesis and screening of alternatives: all the feasible circuit alternatives are generated, the performance of each circuit is estimated using group contribution, circuits are classified based on criteria previously defined, and finally the performance of the most promising circuits is validated by modelling; (iii) improvements and final design: the equipment and operational conditions are determined using modelling and/or experimental tests, local sensitivity analysis is used to identify the most critical variables and reverse simulation is used for modelling. A case study illustrates the proposed methodology. This is a hybrid design method because empirical knowledge, approximate models and designer’s experience are used together with rigorous modelling and testing.

The copper ore in the upper sectors of the El Teniente underground mine in Chile is crushed and ground in the Sewell beneficiation plant and led as pulp through a canal to the flotation plant in Colón. Primary flotation of the Sewell ore is done at pH 4. The ore characteristics, with a high percentage of non-sulphide copper and acid treatment in the grinding stage, have resulted in a natural leaching of this pulp not recoverable in conventional flotation. To recover dissolved Cu in the acid tailings water, one of the technically and economically viable alternatives is to divert the tailings independently to an existing tailings thickener used for solid/liquid separation, particularly for adding special flocculant for this type of pulp. The overflow of recovered water will be driven through a pumping system to the head tank of the solvent extraction plant. The project, currently in development, aims to reduce the specific consumption of flocculant, to increase the percentage of solids in the discharge and to increase water recovery. However, coating the existing thickener, with its concrete base and metal structure, with material resistant to new operating conditions should also be evaluated to avoid any damage caused by acid etching, as it currently operates with alkaline pulp. By determining the optimum operating conditions, a better understanding of the technical feasibility of this new process on an industrial scale may be achieved, and associated capital costs and operational expenditure estimated., The copper ore in the upper sectors of the El Teniente underground mine in Chile is crushed and ground in the Sewell beneficiation plant and led as pulp through a canal to the flotation plant in Colón. Primary flotation of the Sewell ore is done at pH 4. The ore characteristics, with a high percentage of non-sulphide copper and acid treatment in the grinding stage, have resulted in a natural leaching of this pulp not recoverable in conventional flotation. To recover dissolved Cu in the acid tailings water, one of the technically and economically viable alternatives is to divert the tailings independently to an existing tailings thickener used for solid/liquid separation, particularly for adding special flocculant for this type of pulp. The overflow of recovered water will be driven through a pumping system to the head tank of the solvent extraction plant. The project, currently in development, aims to reduce the specific consumption of flocculant, to increase the percentage of solids in the discharge and to increase water recovery. However, coating the existing thickener, with its concrete base and metal structure, with material resistant to new operating conditions should also be evaluated to avoid any damage caused by acid etching, as it currently operates with alkaline pulp. By determining the optimum operating conditions, a better understanding of the technical feasibility of this new process on an industrial scale may be achieved, and associated capital costs and operational expenditure estimated.

Northgate Minerals' Kemess Mine in central British Columbia, Canada, was a large copper-gold porphyry deposit in operation between May 1998 and February 2011. The difficult-to-treat hypogene ore from the East Pit was processed at the end of mine life. That ore had much lower copper (0.12% versus 0.22%) and gold (0.25 versus 0.7 g/ton) head grades with higher pyrite content. Concentrate grade and recovery was poor when it was first processed in early 2005. It was then decided to treat that ore later after first investigating and better understanding the metallurgical behaviour of that ore body. Metallurgical test work and mineralogical liberation analysis showed that a higher concentrate grade was possible when decreasing the regrind product size P80 target from 45 to 21 micrometers. The industrial results confirmed the laboratory test results but at the expense of much higher power draw and media consumption at the regrind mill. In July 2010, the grinding media of the regrind ball mill was converted from 38 mm balls to small grinding media, called Millpebs. As a result, the power draw was decreased from 3 285 to 1 600 kW with achievement of a slightly finer grind. The final copper grade was improved by 1.2% with an increase of 2.6% in copper and 3.6% in gold recoveries of the cleaner flotation circuit. The net benefit was about Can, Northgate Minerals' Kemess Mine in central British Columbia, Canada, was a large copper-gold porphyry deposit in operation between May 1998 and February 2011. The difficult-to-treat hypogene ore from the East Pit was processed at the end of mine life. That ore had much lower copper (0.12% versus 0.22%) and gold (0.25 versus 0.7 g/ton) head grades with higher pyrite content. Concentrate grade and recovery was poor when it was first processed in early 2005. It was then decided to treat that ore later after first investigating and better understanding the metallurgical behaviour of that ore body. Metallurgical test work and mineralogical liberation analysis showed that a higher concentrate grade was possible when decreasing the regrind product size P80 target from 45 to 21 micrometers. The industrial results confirmed the laboratory test results but at the expense of much higher power draw and media consumption at the regrind mill. In July 2010, the grinding media of the regrind ball mill was converted from 38 mm balls to small grinding media, called Millpebs. As a result, the power draw was decreased from 3 285 to 1 600 kW with achievement of a slightly finer grind. The final copper grade was improved by 1.2% with an increase of 2.6% in copper and 3.6% in gold recoveries of the cleaner flotation circuit. The net benefit was about Can

The deportment of chalcocite (Cu2S) and heazlewoodite (Ni3S2) in Vale's Bessemer matte flotation circuit in Sudbury, Ontario, was investigated both in the plant and in laboratory flotation experiments. Insufficient selectivity was reported due to Cu2S not being entirely recovered in the rougher/scavengers and Ni3S2, depressed in the process, contaminating the flotation concentrate. It was found that Cu2S misreporting to the flotation tailings is mainly ultrafine, smaller than 6 micrometers, suggesting that surface oxidation and the low collision efficiency of fine particles in flotation are the possible causes of Cu2S losses. By contrast, Ni3S2 misreporting to the flotation concentrate is mainly in the intermediate size fractions from 6 to 38 micrometers and less than 40% of the heazlewoodite is in composites with chalcocite, the majority being liberated. Particle surface analysis by means of time-of-flight secondary ion mass spectrometry revealed that Ni3S2 in the concentrate is Cu-activated, and that the diphenyl guanidine collector adsorbs on the exposed Cu on the mineral surface. The role of pulp electrochemistry was also investigated, showing that the pulp potential in primary grinding is critical for selectivity. Strategies to reduce Cu activation of Ni3S2 in the process are suggested, as well as modifications to the current circuit flowsheet for improved plant performance., The deportment of chalcocite (Cu2S) and heazlewoodite (Ni3S2) in Vale's Bessemer matte flotation circuit in Sudbury, Ontario, was investigated both in the plant and in laboratory flotation experiments. Insufficient selectivity was reported due to Cu2S not being entirely recovered in the rougher/scavengers and Ni3S2, depressed in the process, contaminating the flotation concentrate. It was found that Cu2S misreporting to the flotation tailings is mainly ultrafine, smaller than 6 micrometers, suggesting that surface oxidation and the low collision efficiency of fine particles in flotation are the possible causes of Cu2S losses. By contrast, Ni3S2 misreporting to the flotation concentrate is mainly in the intermediate size fractions from 6 to 38 micrometers and less than 40% of the heazlewoodite is in composites with chalcocite, the majority being liberated. Particle surface analysis by means of time-of-flight secondary ion mass spectrometry revealed that Ni3S2 in the concentrate is Cu-activated, and that the diphenyl guanidine collector adsorbs on the exposed Cu on the mineral surface. The role of pulp electrochemistry was also investigated, showing that the pulp potential in primary grinding is critical for selectivity. Strategies to reduce Cu activation of Ni3S2 in the process are suggested, as well as modifications to the current circuit flowsheet for improved plant performance.

One of Vale’s sponsored research projects is aimed at the implementation of a practical scale-up procedure for flotation. This work presents the results of characterising flotation kinetic parameters with respect to particle composition. A special type of iron ore known as compact itabirite was chosen because it consists essentially of two phases, quartz and haematite, which in turn have very distinct densities. This type of ore is amenable to ionic quartz flotation and produces concentrates within industrial specification. A series of standard kinetic tests were carried out in order to produce enough material in each flotation time for size analysis and mineralogical characterisation. Mineral liberation analysis was carried out for 40 narrow-sized particle samples with known flotation time. This produces grade distributions conditional on size and flotation time, with the traditional twelve grade classes ranging from completely liberated quartz to completely liberated haematite. The calculations involve the determination of recoveries for particles of the same type, i.e. size and composition, at each flotation time. Results show that the flotation rate of the particles does not vary much with composition and that the fraction of floatable particles in each grade class controls the separation action. Particle size also has an important role in the floatable phase fraction., One of Vale’s sponsored research projects is aimed at the implementation of a practical scale-up procedure for flotation. This work presents the results of characterising flotation kinetic parameters with respect to particle composition. A special type of iron ore known as compact itabirite was chosen because it consists essentially of two phases, quartz and haematite, which in turn have very distinct densities. This type of ore is amenable to ionic quartz flotation and produces concentrates within industrial specification. A series of standard kinetic tests were carried out in order to produce enough material in each flotation time for size analysis and mineralogical characterisation. Mineral liberation analysis was carried out for 40 narrow-sized particle samples with known flotation time. This produces grade distributions conditional on size and flotation time, with the traditional twelve grade classes ranging from completely liberated quartz to completely liberated haematite. The calculations involve the determination of recoveries for particles of the same type, i.e. size and composition, at each flotation time. Results show that the flotation rate of the particles does not vary much with composition and that the fraction of floatable particles in each grade class controls the separation action. Particle size also has an important role in the floatable phase fraction.

The mechanism of sphalerite activation with Pb(II) was investigated under conditions similar to those of flotation circuits for complex sulphide minerals. The contact angle technique was used, which allows fairly robust measurements of hydrophobisation of sphalerite crystals under the chemical conditions of interest. The results show that activation is apparently due to adsorption of lead hydroxo-complexes (PbOH+) on the sphalerite surface, followed by precipitation of Pb(OH)2 or PbS, both of which form stable hydrophobic compounds with xanthate. The presence of dissolved oxygen enhances sphalerite hydrophobicity, presumably due to the catalytic effect that Pb(II) plays on the oxidation of S2- with oxygen to produce elemental sulphur or polysulphide. Sphalerite activation with Pb(II) occurs regardless of whether lead is intentionally added to the flotation slurry or generated in situ by galena oxidation with the dissolved oxygen present in the pulp. EDTA extraction of the oxidised lead on galena showed that fairly mild oxidation is sufficient to accomplish substantial activation of sphalerite. Finally, based on thermodynamic analysis of the system, it was experimentally corroborated that addition of HS- is an effective way of controlling sphalerite activation, presumably by fixing the Pb(II) species on the surface of galena particles, preventing its migration to sphalerite particles., The mechanism of sphalerite activation with Pb(II) was investigated under conditions similar to those of flotation circuits for complex sulphide minerals. The contact angle technique was used, which allows fairly robust measurements of hydrophobisation of sphalerite crystals under the chemical conditions of interest. The results show that activation is apparently due to adsorption of lead hydroxo-complexes (PbOH+) on the sphalerite surface, followed by precipitation of Pb(OH)2 or PbS, both of which form stable hydrophobic compounds with xanthate. The presence of dissolved oxygen enhances sphalerite hydrophobicity, presumably due to the catalytic effect that Pb(II) plays on the oxidation of S2- with oxygen to produce elemental sulphur or polysulphide. Sphalerite activation with Pb(II) occurs regardless of whether lead is intentionally added to the flotation slurry or generated in situ by galena oxidation with the dissolved oxygen present in the pulp. EDTA extraction of the oxidised lead on galena showed that fairly mild oxidation is sufficient to accomplish substantial activation of sphalerite. Finally, based on thermodynamic analysis of the system, it was experimentally corroborated that addition of HS- is an effective way of controlling sphalerite activation, presumably by fixing the Pb(II) species on the surface of galena particles, preventing its migration to sphalerite particles.

This work shows results of bench-scale flotation studies for nickel ore. Studies consisted of mineralogical characterisation by scanning electron microscopy and chemical analysis by X-ray fluorescence. These studies revealed the presence of minerals such as magnesiohornblende, pyroxmangite, ferrosilite, talc, pyrrhotite, quartz and pentlandite. The Ni grade in the ore was approximately 0.80%. Significant factors influencing the recovery of Ni in flotation studies were identified using a fractional factorial design of 32 experiments. Six factors were investigated: collector type (Hostaflot E501 and Hostaflot M92), collector concentration (g/t), frother type (Flotanol D25 and Montanol), frother concentration (g/t), dispersant concentration (Dispersol 589 (g/t) and depressant concentration (carboxymethyl cellulose (g/t). The results showed that only the type of collector and the type of frother significantly influence the response variables. The recovery of nickel obtained in the rougher cell was approximately 84% with 2.1% Ni and 13.8% MgO., This work shows results of bench-scale flotation studies for nickel ore. Studies consisted of mineralogical characterisation by scanning electron microscopy and chemical analysis by X-ray fluorescence. These studies revealed the presence of minerals such as magnesiohornblende, pyroxmangite, ferrosilite, talc, pyrrhotite, quartz and pentlandite. The Ni grade in the ore was approximately 0.80%. Significant factors influencing the recovery of Ni in flotation studies were identified using a fractional factorial design of 32 experiments. Six factors were investigated: collector type (Hostaflot E501 and Hostaflot M92), collector concentration (g/t), frother type (Flotanol D25 and Montanol), frother concentration (g/t), dispersant concentration (Dispersol 589 (g/t) and depressant concentration (carboxymethyl cellulose (g/t). The results showed that only the type of collector and the type of frother significantly influence the response variables. The recovery of nickel obtained in the rougher cell was approximately 84% with 2.1% Ni and 13.8% MgO.

The advent of direct measurement of required pinion energy for semi-autogenous grinding mills has created much controversy as to which method is best for the design of new mills. Direct measurement using open technology is preferable to simulated solutions because the client can check the design themselves. Many closed and private technologies make checking of a particular design difficult, if not impossible, without the purchase of expensive software, if available. Industry consensus among professional engineers demands that new designs be checked by at least two mainstream methods to provide adequate assurance that the design offered will work and will process the design tonnage at desired product size every day in the new plant. Too much faith in simulated design and lack of experience in its application can also result in increased start-up times due to the mill sizing having been inferred and not calculated. The paper examines the use of SAGDesign technology as a valid method for determining the chamber size and power required for new autogenou-ball, semiautogenous-ball and semiautogenous-ball-crusher grinding circuits. Test accuracy, benchmark results and database summary information are included., The advent of direct measurement of required pinion energy for semi-autogenous grinding mills has created much controversy as to which method is best for the design of new mills. Direct measurement using open technology is preferable to simulated solutions because the client can check the design themselves. Many closed and private technologies make checking of a particular design difficult, if not impossible, without the purchase of expensive software, if available. Industry consensus among professional engineers demands that new designs be checked by at least two mainstream methods to provide adequate assurance that the design offered will work and will process the design tonnage at desired product size every day in the new plant. Too much faith in simulated design and lack of experience in its application can also result in increased start-up times due to the mill sizing having been inferred and not calculated. The paper examines the use of SAGDesign technology as a valid method for determining the chamber size and power required for new autogenou-ball, semiautogenous-ball and semiautogenous-ball-crusher grinding circuits. Test accuracy, benchmark results and database summary information are included.

Applying greater power to flotation slurries yields significant improvements in fine particle recovery; similarly lowering the absorbed power results in increased coarse particle recovery. An improvement in the kinetics of fine and coarse size classes, provided there is no adverse metallurgical influence on intermediate size ranges, is beneficial to the overall recovery response. The practical implementation of this concept was recently introduced by FLSmidth. In principle, it decouples flotation regimes where fine and coarse particles are preferentially recovered. In the case of naturally aspirated machines (WEMCO) it is called Hybrid Energy Flotation technology, in the case of forced air machines (Dorr-Oliver) it is called Hybrid Mechanism Flotation and incorporates three sections regardless of the type of the machine. In the first section, at the beginning of the row, where flotation is froth-phase limited and operational and set-up parameters have a limited influence on the recovery, standard flotation machines are used. In the second section higher-powered flotation machines are used to improve fine particle recovery. In the third section lower-powered flotation machines enhance coarse particle recovery. The placement of low- and high-energy machines in a selected row is application specific., Applying greater power to flotation slurries yields significant improvements in fine particle recovery; similarly lowering the absorbed power results in increased coarse particle recovery. An improvement in the kinetics of fine and coarse size classes, provided there is no adverse metallurgical influence on intermediate size ranges, is beneficial to the overall recovery response. The practical implementation of this concept was recently introduced by FLSmidth. In principle, it decouples flotation regimes where fine and coarse particles are preferentially recovered. In the case of naturally aspirated machines (WEMCO) it is called Hybrid Energy Flotation technology, in the case of forced air machines (Dorr-Oliver) it is called Hybrid Mechanism Flotation and incorporates three sections regardless of the type of the machine. In the first section, at the beginning of the row, where flotation is froth-phase limited and operational and set-up parameters have a limited influence on the recovery, standard flotation machines are used. In the second section higher-powered flotation machines are used to improve fine particle recovery. In the third section lower-powered flotation machines enhance coarse particle recovery. The placement of low- and high-energy machines in a selected row is application specific.

A typical mobile crushing plant is a challenging target for automatic controls due to lack of instrumentation for plant-wide controls. A single crusher produces quite a lot information on its own operation such as power, pressure and crusher level. It also may be connected to a feeder of the same crushing unit for regulating the material level in the crusher. However, the units lack coordinating and supervisory controls that would link all the crushing units together and stabilise the material flows in all phases. The mobility of crushing units poses some issues for automatic controls. First of all, the plant layout, that is, the order of the crushing units operating in sequence, may vary from time to time. Consequently, the automatic feed controls must be agile and capable of reorganising themselves according to the plant layout at a time. Second, mobile crushing units change more or less frequently their on-site positions. Process interruptions due to plant layout and position changes require fast and robust automatic start-ups and shutdowns for reducing downtime and production losses. This paper introduces feed control design for a three-stage mobile crushing plant allowing four different plant layouts. Each design is given in detail and a simulated example is shown for one of the layouts., A typical mobile crushing plant is a challenging target for automatic controls due to lack of instrumentation for plant-wide controls. A single crusher produces quite a lot information on its own operation such as power, pressure and crusher level. It also may be connected to a feeder of the same crushing unit for regulating the material level in the crusher. However, the units lack coordinating and supervisory controls that would link all the crushing units together and stabilise the material flows in all phases. The mobility of crushing units poses some issues for automatic controls. First of all, the plant layout, that is, the order of the crushing units operating in sequence, may vary from time to time. Consequently, the automatic feed controls must be agile and capable of reorganising themselves according to the plant layout at a time. Second, mobile crushing units change more or less frequently their on-site positions. Process interruptions due to plant layout and position changes require fast and robust automatic start-ups and shutdowns for reducing downtime and production losses. This paper introduces feed control design for a three-stage mobile crushing plant allowing four different plant layouts. Each design is given in detail and a simulated example is shown for one of the layouts.

In the reverse flotation of iron ore, it is common to use starch as a depressant. The addition of water and sodium hydroxide is usually required for the preparation of starch. This study evaluated the feasibility of using the optical refractometer for starch quantification in solutions used as depressants. These solutions are prepared with cassava meal and corn starch, and its gelatinisation process is carried out with the addition of sodium hydroxide. The solutions prepared contained different starch/caustic soda ratios. The results indicated that the concentration readings from the refractometer do not correspond to the actual value calculated. This is due to the presence of two solutes (starch and sodium hydroxide) to one solvent (water). The concentration of starch only represents the actual one used for the preparation of the solution when all starch was gelatinised by sodium hydroxide. Thus, during the process of measuring the concentration of the solution for the other starch/sodium hydroxide proportions, the refractometer was not only measuring the quantity of starch, but also the quantity of caustic soda., In the reverse flotation of iron ore, it is common to use starch as a depressant. The addition of water and sodium hydroxide is usually required for the preparation of starch. This study evaluated the feasibility of using the optical refractometer for starch quantification in solutions used as depressants. These solutions are prepared with cassava meal and corn starch, and its gelatinisation process is carried out with the addition of sodium hydroxide. The solutions prepared contained different starch/caustic soda ratios. The results indicated that the concentration readings from the refractometer do not correspond to the actual value calculated. This is due to the presence of two solutes (starch and sodium hydroxide) to one solvent (water). The concentration of starch only represents the actual one used for the preparation of the solution when all starch was gelatinised by sodium hydroxide. Thus, during the process of measuring the concentration of the solution for the other starch/sodium hydroxide proportions, the refractometer was not only measuring the quantity of starch, but also the quantity of caustic soda.

Six Sigma, widely adopted as a discipline in a variety of industries including mining, is the data-driven problem solving approach supported by a handful of powerful statistical tools in order to reduce variation through a continuous process of improvement. The methodology consists of five essential steps: defining the problem, measuring the current state, analysing the root causes and improving and controlling the achieved process results. This paper proposes a Six Sigma methodology based on data mining for effectively and efficiently processing massive data driving the area of milling. As a strategy in Minera Fresnillo, Fresnillo PLC processes lower-grade ores of silver and has drawn up the production schedules required to increase production capacity from 7 500 to 8 000 tons, but this increase will coarsen the flotation feed which will directly affect silver recovery. With the methodology, an increase in the recovery of silver in lead concentrate of 85.29-86.03% was achieved with an optimum particle size of the ore fed to flotation of 90% finer than 200 mesh at a rate of 8 000 t by day, making a profit of US, Six Sigma, widely adopted as a discipline in a variety of industries including mining, is the data-driven problem solving approach supported by a handful of powerful statistical tools in order to reduce variation through a continuous process of improvement. The methodology consists of five essential steps: defining the problem, measuring the current state, analysing the root causes and improving and controlling the achieved process results. This paper proposes a Six Sigma methodology based on data mining for effectively and efficiently processing massive data driving the area of milling. As a strategy in Minera Fresnillo, Fresnillo PLC processes lower-grade ores of silver and has drawn up the production schedules required to increase production capacity from 7 500 to 8 000 tons, but this increase will coarsen the flotation feed which will directly affect silver recovery. With the methodology, an increase in the recovery of silver in lead concentrate of 85.29-86.03% was achieved with an optimum particle size of the ore fed to flotation of 90% finer than 200 mesh at a rate of 8 000 t by day, making a profit of US

Brazilian pegmatite lithium mineral production is very modest compared with the world Li mining industry. Flotation is scarcely used in Brazil, mainly due to difficulties inherent to the process of selectively separating spodumene from other silicates in pegmatite formations, but it is an interesting alternative to the processes currently used. The fine particle fraction, smaller than about 1 mm, in the gravity beneficiation method is currently discarded as tailings. Worldwide literature shows few fundamental studies related to spodumene flotation. Considering the main minerals associated with spodumene in pegmatites are microcline, muscovite and quartz, the present work aimed at investigating amine adsorption on the surface of such minerals. Microflotation tests and zeta potential measurements were used and results showed evidence that differences in the floatability of minerals are directly related to their surface charges. The primary adsorption mechanism can be attributed to electrostatic forces. The results also show that spodumene can be selectively separated from the studied minerals through microflotation at pH 5. The investigation indicated that there is a great potential in using a spodumene flotation route based on the use of amine as a collector., Brazilian pegmatite lithium mineral production is very modest compared with the world Li mining industry. Flotation is scarcely used in Brazil, mainly due to difficulties inherent to the process of selectively separating spodumene from other silicates in pegmatite formations, but it is an interesting alternative to the processes currently used. The fine particle fraction, smaller than about 1 mm, in the gravity beneficiation method is currently discarded as tailings. Worldwide literature shows few fundamental studies related to spodumene flotation. Considering the main minerals associated with spodumene in pegmatites are microcline, muscovite and quartz, the present work aimed at investigating amine adsorption on the surface of such minerals. Microflotation tests and zeta potential measurements were used and results showed evidence that differences in the floatability of minerals are directly related to their surface charges. The primary adsorption mechanism can be attributed to electrostatic forces. The results also show that spodumene can be selectively separated from the studied minerals through microflotation at pH 5. The investigation indicated that there is a great potential in using a spodumene flotation route based on the use of amine as a collector.

Comminution generally leads to a distribution of different sized particles, which inevitably means that oversize and/or undersize fractions are produced together with the desired/most valued fractions. Consequently, it is impossible to determine a single parameter that would intuitively describe the optimum operation point. As a result, the optimality criterion has to be formulated as a function of case-specific production targets in order to find the optimum balance between production rates of products and by-products, energy consumption and wear of crushing liners. Doing this enables the circuit operating profit to be maximised with respect to time. This paper introduces a model-based optimisation method for optimum selection of operating points in cone crushers, formulated as a multiobjective profit function with respect to the case-specific production objectives. The production objectives in profit function can then be weighted according to the current demands, which are set either by the market or the downstream operations, in order to provide optimum operating conditions for the given situation. The result is an intuitive 3D description of the entire cone-crusher operating range as a function of financial profit with respect to the crusher manipulated variables; that is, closed-side setting and eccentric speed., Comminution generally leads to a distribution of different sized particles, which inevitably means that oversize and/or undersize fractions are produced together with the desired/most valued fractions. Consequently, it is impossible to determine a single parameter that would intuitively describe the optimum operation point. As a result, the optimality criterion has to be formulated as a function of case-specific production targets in order to find the optimum balance between production rates of products and by-products, energy consumption and wear of crushing liners. Doing this enables the circuit operating profit to be maximised with respect to time. This paper introduces a model-based optimisation method for optimum selection of operating points in cone crushers, formulated as a multiobjective profit function with respect to the case-specific production objectives. The production objectives in profit function can then be weighted according to the current demands, which are set either by the market or the downstream operations, in order to provide optimum operating conditions for the given situation. The result is an intuitive 3D description of the entire cone-crusher operating range as a function of financial profit with respect to the crusher manipulated variables; that is, closed-side setting and eccentric speed.

The use of online pulp analyser assaying to determine the various process streams and their integration into flotation expert systems is the key to improving and optimising the processes of mineral beneficiation. Reliable analysers, with only small differences in their measurements from those tested in the laboratory, and robust strategies in expert systems ensure stability in the process and maximise the metallurgical results. The project was developed at Minera Fresnillo, where through the application of Six Sigma in the silver-lead circuit flotation process differences in the measurements of online pulp analysers were reduced through the improvement of models and the development and implementation of a predictive equation for the optimum level of Ag in Pb concentrate. This optimum level is according to the grade of Ag and Pb in the head feed to the circuit, and combined with the logic of the expert system, the recovery of Ag in the concentrate was improved, impacting economic recovery. The latest results obtained in the project were a 1% increase in Ag recovery to the Pb concentrate, from 84% to 85%, and an increase from 13 300 to 14 000 kg in the grade of Ag in the Pb concentrate, leading to an economic profit of US, The use of online pulp analyser assaying to determine the various process streams and their integration into flotation expert systems is the key to improving and optimising the processes of mineral beneficiation. Reliable analysers, with only small differences in their measurements from those tested in the laboratory, and robust strategies in expert systems ensure stability in the process and maximise the metallurgical results. The project was developed at Minera Fresnillo, where through the application of Six Sigma in the silver-lead circuit flotation process differences in the measurements of online pulp analysers were reduced through the improvement of models and the development and implementation of a predictive equation for the optimum level of Ag in Pb concentrate. This optimum level is according to the grade of Ag and Pb in the head feed to the circuit, and combined with the logic of the expert system, the recovery of Ag in the concentrate was improved, impacting economic recovery. The latest results obtained in the project were a 1% increase in Ag recovery to the Pb concentrate, from 84% to 85%, and an increase from 13 300 to 14 000 kg in the grade of Ag in the Pb concentrate, leading to an economic profit of US

Vale, the second largest mining company in the world, has instigated a detailed study for the utilisation of Vertimills as an alternative to conventional tube ball mills in pellet feed grinding applications. The Vertimill has been used successfully in the past few decades, mainly for fine grinding applications in which it has been reported as 35% more efficient than the conventional tube ball mill. The main objective of this work is to establish a scale-up procedure for the Vertimill, currently based on acquired experiences in similar operations, on laboratory batch ball-mill tests and on pilot-scale tests. Results are presented from a series of bench-scale tests for determining population-balance model parameters. The scale-up procedure is based on the Herbst-Fuerstenau scale-up procedure as implemented in Vale's laboratories. Simulations using these laboratory parameters were validated by scaling up a Vertimill pilot-scale test carried out at Metso's pilot-plant facility in York, Pennsylvania., Vale, the second largest mining company in the world, has instigated a detailed study for the utilisation of Vertimills as an alternative to conventional tube ball mills in pellet feed grinding applications. The Vertimill has been used successfully in the past few decades, mainly for fine grinding applications in which it has been reported as 35% more efficient than the conventional tube ball mill. The main objective of this work is to establish a scale-up procedure for the Vertimill, currently based on acquired experiences in similar operations, on laboratory batch ball-mill tests and on pilot-scale tests. Results are presented from a series of bench-scale tests for determining population-balance model parameters. The scale-up procedure is based on the Herbst-Fuerstenau scale-up procedure as implemented in Vale's laboratories. Simulations using these laboratory parameters were validated by scaling up a Vertimill pilot-scale test carried out at Metso's pilot-plant facility in York, Pennsylvania.

Hukki's law of 1962 is an empirical law which does not take into account several events of energy loss during the mineral fragmentation processes. Since experimental results are very difficult to obtain for a large range of fragment sizes, verification of the law is very difficult. A relationship was proposed some decades ago between the fractures and fragmentation processes on the one hand and fractal geometry on the other, empirical laws known in this context showing basic features of fractal geometry, mainly self-affinity and power law behaviour. In this paper a model was developed to simulate the fragmentation process and to check the relation between energy consumption and fragment sizes. The model is represented on a regular lattice where links represent pathways for fracture processes. The energy of fragmentation events is modelled by a probability distribution function. In the proposed model there is no mass loss and the fracture propagation occurs as self-avoiding random walks on the regular lattice. (Authors.), Hukki's law of 1962 is an empirical law which does not take into account several events of energy loss during the mineral fragmentation processes. Since experimental results are very difficult to obtain for a large range of fragment sizes, verification of the law is very difficult. A relationship was proposed some decades ago between the fractures and fragmentation processes on the one hand and fractal geometry on the other, empirical laws known in this context showing basic features of fractal geometry, mainly self-affinity and power law behaviour. In this paper a model was developed to simulate the fragmentation process and to check the relation between energy consumption and fragment sizes. The model is represented on a regular lattice where links represent pathways for fracture processes. The energy of fragmentation events is modelled by a probability distribution function. In the proposed model there is no mass loss and the fracture propagation occurs as self-avoiding random walks on the regular lattice. (Authors.)

The 135 000 tpd capacity Cu and Mo concentration plant at Doña Inés de Collahuasi, a copper mining company in Chile, uses underground water from salt flats. The investigation aimed to reduce sulphate and chloride content of the salt water by adding lime and treated tailings water. Thickening tests were also carried out on final tailings where 14% dilution was reached using treated water. Results show that the agitation speed is an important factor for the removal of chlorides but not sulphates. The lime dosage is a significant variable since, in addition to raising the pH, lime reacts with the dissolved impurities that precipitate as metallic hydroxides and calcium sulphate. The reaction of salt water with lime and tailings is more efficient for removal of chlorides than for sulphates, but this mixture with final tailings, at 30% solid contents, produces a thickener overflow of better quality concerning sulphate and chloride content. For a volumetric ratio of 0.1, the global decrease of dissolved sulphate and chloride concentrations is over 50%. By adding flocculant, the final concentrations of sulphates and chlorides can be reduced from 5 to 1.4 g/l and from 35 to 10 g/l, respectively. The effect on flotation performance was also studied. Rougher Cu and Mo recoveries were increased to 0.5% and 2% respectively by using treated water., The 135 000 tpd capacity Cu and Mo concentration plant at Doña Inés de Collahuasi, a copper mining company in Chile, uses underground water from salt flats. The investigation aimed to reduce sulphate and chloride content of the salt water by adding lime and treated tailings water. Thickening tests were also carried out on final tailings where 14% dilution was reached using treated water. Results show that the agitation speed is an important factor for the removal of chlorides but not sulphates. The lime dosage is a significant variable since, in addition to raising the pH, lime reacts with the dissolved impurities that precipitate as metallic hydroxides and calcium sulphate. The reaction of salt water with lime and tailings is more efficient for removal of chlorides than for sulphates, but this mixture with final tailings, at 30% solid contents, produces a thickener overflow of better quality concerning sulphate and chloride content. For a volumetric ratio of 0.1, the global decrease of dissolved sulphate and chloride concentrations is over 50%. By adding flocculant, the final concentrations of sulphates and chlorides can be reduced from 5 to 1.4 g/l and from 35 to 10 g/l, respectively. The effect on flotation performance was also studied. Rougher Cu and Mo recoveries were increased to 0.5% and 2% respectively by using treated water.

Knowledge of a material's behaviour allows a better understanding of important phenomena which occur inside silos, such as discharge; previous studies to simulate the flow of granular material in silos have included use of discrete-element, molecular-dynamic and event-driven modelling of particles as elastic or rigid. The computational model presented in this paper has been developed using the technique known as cellular automata to simulate a granular dense but porous package inside a silo, generating a granular material flow through the silo discharge outlet. The data generated by simulation were validated with the help of an experimental low-cost device in acrylic, a bi-dimensional silo that allows the setting up of outlet inclination. After a linear regression had been performed, the results showed an acceptable fit between the computational model and the experimental data., Knowledge of a material's behaviour allows a better understanding of important phenomena which occur inside silos, such as discharge; previous studies to simulate the flow of granular material in silos have included use of discrete-element, molecular-dynamic and event-driven modelling of particles as elastic or rigid. The computational model presented in this paper has been developed using the technique known as cellular automata to simulate a granular dense but porous package inside a silo, generating a granular material flow through the silo discharge outlet. The data generated by simulation were validated with the help of an experimental low-cost device in acrylic, a bi-dimensional silo that allows the setting up of outlet inclination. After a linear regression had been performed, the results showed an acceptable fit between the computational model and the experimental data.

A system of continuous SAG grinding laboratory tests has been designed and manufactured, and procedures for carrying out tests to evaluate changes in the geometry of the mill and discharge system as well as operational parameters have been developed. Based on tests conducted in this system and considering a simplified SAG model presented in other work, the influence of variations in operational parameters on the phenomena involved are analysed and represented in the model used to identify important features for understanding processes. The analysis includes the effect of variations in speed, ball fill level and solids percentage on the size reduction processes, power consumption, internal classification and discharge processes, considering the additional determination of the internal characteristics of load produced in detail, including size distribution and solid percentage.Changes in effective fracture rate, Ki, are globally analysed with each of the previous parameters and the differences in specific energy consumption obtained. Full size distribution curves of internal load obtained from laboratory tests are presented and the internal classification process analysed as a function of operational parameters through Ci estimation., A system of continuous SAG grinding laboratory tests has been designed and manufactured, and procedures for carrying out tests to evaluate changes in the geometry of the mill and discharge system as well as operational parameters have been developed. Based on tests conducted in this system and considering a simplified SAG model presented in other work, the influence of variations in operational parameters on the phenomena involved are analysed and represented in the model used to identify important features for understanding processes. The analysis includes the effect of variations in speed, ball fill level and solids percentage on the size reduction processes, power consumption, internal classification and discharge processes, considering the additional determination of the internal characteristics of load produced in detail, including size distribution and solid percentage.Changes in effective fracture rate, Ki, are globally analysed with each of the previous parameters and the differences in specific energy consumption obtained. Full size distribution curves of internal load obtained from laboratory tests are presented and the internal classification process analysed as a function of operational parameters through Ci estimation.

The mining industry is currently facing significant sustainability challenges in terms of dealing with difficult-to-process low-grade resources. Different strategies are discussed for dealing with gangue sulphides, such as pyrite, and penalty elements and the most efficient alternative is proposed. New products and application technologies have been developed in the past years for implementing these strategies as dictated by the particular needs of a given plant. However, in recognition of the growing interest in meeting the sustainability challenges, Cytec Research and Development has focused on the development of greener products (collectors, modifiers and frothers) and processes using the Flotation Matrix 100 approach. An overview of these new developments is given using case studies involving the application of selective collectors and polymeric modifiers including the newer, greener chemistries. The results presented demonstrate that flotation is more efficient with highly selective collectors in the rougher stage than collectors with low or medium selectivity in conjunction with mostly high doses of depressants, especially organic depressants, which are shown to be inefficient and potentially to affect the recovery of both the main sulphide product and secondary products such as molybdenum and gold., The mining industry is currently facing significant sustainability challenges in terms of dealing with difficult-to-process low-grade resources. Different strategies are discussed for dealing with gangue sulphides, such as pyrite, and penalty elements and the most efficient alternative is proposed. New products and application technologies have been developed in the past years for implementing these strategies as dictated by the particular needs of a given plant. However, in recognition of the growing interest in meeting the sustainability challenges, Cytec Research and Development has focused on the development of greener products (collectors, modifiers and frothers) and processes using the Flotation Matrix 100 approach. An overview of these new developments is given using case studies involving the application of selective collectors and polymeric modifiers including the newer, greener chemistries. The results presented demonstrate that flotation is more efficient with highly selective collectors in the rougher stage than collectors with low or medium selectivity in conjunction with mostly high doses of depressants, especially organic depressants, which are shown to be inefficient and potentially to affect the recovery of both the main sulphide product and secondary products such as molybdenum and gold.

Doña Inés de Collahuasi, a copper mining company in Chile’s First Region, Tarapacá, includes a concentration plant with a processing capacity of 135 000 tdp located at 4 320 m above sea-level. The flotation stage of the Cu and Mo concentrator consists of a rougher flotation circuit equipped with both conventional self-aerated cells and injected-air cells. The cleaning circuit operates in two stages; the first uses conventional self-aerated cells and the second cleaning stage is performed in columns. Extensive biographical information supports the correlation between the bubble size and the recovery of sulphide minerals. Based on this, the company decided to carry out an evaluation to determine the effect of geographical altitude on the bubble size. Measurements were done using the same laboratory-scale flotation cell, the same standard frother, and the same technique for bubble size measurement. The flotation cell, installed on a vehicle to allow measurements at different altitudes, was set up to measure and control the air flow online, with the option of being operated either with self-aerated or injected air. It was concluded that the bubble size generated in self-aerated cells is not affected either by agitation speed or geographical altitude; the air flow measured inside the self-aerated cell is not affected by geographical altitude; bubble size is not affected by the way in which the airflow is supplied; the smallest bubble size measured was using the self-aerated flotation cell., Doña Inés de Collahuasi, a copper mining company in Chile’s First Region, Tarapacá, includes a concentration plant with a processing capacity of 135 000 tdp located at 4 320 m above sea-level. The flotation stage of the Cu and Mo concentrator consists of a rougher flotation circuit equipped with both conventional self-aerated cells and injected-air cells. The cleaning circuit operates in two stages; the first uses conventional self-aerated cells and the second cleaning stage is performed in columns. Extensive biographical information supports the correlation between the bubble size and the recovery of sulphide minerals. Based on this, the company decided to carry out an evaluation to determine the effect of geographical altitude on the bubble size. Measurements were done using the same laboratory-scale flotation cell, the same standard frother, and the same technique for bubble size measurement. The flotation cell, installed on a vehicle to allow measurements at different altitudes, was set up to measure and control the air flow online, with the option of being operated either with self-aerated or injected air. It was concluded that the bubble size generated in self-aerated cells is not affected either by agitation speed or geographical altitude; the air flow measured inside the self-aerated cell is not affected by geographical altitude; bubble size is not affected by the way in which the airflow is supplied; the smallest bubble size measured was using the self-aerated flotation cell.

Currently there are many modelling methods for semi-autogenous grinding systems based on phenomenological or empirical approaches, both of which are considered. It is important to discuss the process of mass transport inside the mill, in mathematical representation usually referred to as mass transport law, and whether it properly manages the phenomenon. In addition, a representation must be developed for the effect of the discharge system on the overall process. This document presents an update of the approach developed by Magne (1998) based on phenomenological principles and groups' characteristic phenomena of the size reduction process under a single vector in a simplified form that identifies the grinding process efficiency, called Ki. The internal classification process, which considers grate and discharge system is represented by a characteristic vector, Ci, representing classification efficiency and discharge, without distinguishing between phenomena such as recirculation of pulp and retention of material in the discharge process. These processes are represented by the mass transport function, which is a direct function of the internal load of the mill and the actual flow towards the discharge grate. The model is used to represent the behaviour of the size reduction process in SAG mills through a simple calculation routine applied to continuous tests on a laboratory-scale mill., Currently there are many modelling methods for semi-autogenous grinding systems based on phenomenological or empirical approaches, both of which are considered. It is important to discuss the process of mass transport inside the mill, in mathematical representation usually referred to as mass transport law, and whether it properly manages the phenomenon. In addition, a representation must be developed for the effect of the discharge system on the overall process. This document presents an update of the approach developed by Magne (1998) based on phenomenological principles and groups' characteristic phenomena of the size reduction process under a single vector in a simplified form that identifies the grinding process efficiency, called Ki. The internal classification process, which considers grate and discharge system is represented by a characteristic vector, Ci, representing classification efficiency and discharge, without distinguishing between phenomena such as recirculation of pulp and retention of material in the discharge process. These processes are represented by the mass transport function, which is a direct function of the internal load of the mill and the actual flow towards the discharge grate. The model is used to represent the behaviour of the size reduction process in SAG mills through a simple calculation routine applied to continuous tests on a laboratory-scale mill.