Search Documents

By R. D. Alorro, B. Tadesse, E. Aghaei

Cinnabar (HgS) and native mercury (Hg) are the most common naturally occurring forms of mercury found in most gold ores subsequently released into the environment during processing. Mercury and its compounds are extremely toxic and recognised as serious health and environmental hazards. This research investigates the possibility of mercury (Hg) removal from waste solutions from gold processing using iron-aluminium (Fe-Al) bimetallic particles. The synthesis of Fe-Al bimetallic particles using zero-valent aluminium and ferric chloride as precursors were performed following the optimisation of influencing factors. The presence of both metals in the prepared bimetals was confirmed by XRD analysis. The experiments for mercury removal by bimetals were designed using the central composite design and Minitab software based on two variables including contact time and adsorbent dosage. The best result of Hg(II) removal (>90%) was achieved using 4 g/L bimetals and contact time of 75 min. However, the statistical analysis of results indicated that these parameters have only a minor contribution to the removal of Hg(II) by Fe-Al bimetals. Higher bimetal dosage and contact time resulted in a higher Fe dissolution from the Al surface which can be attributed to the presence of nitrate (# $) ions in the solution. However, adding 0.1 M NaCl to the Hg solution had a positive impact on Hg removal by improving the conductivity and electron transfer within the solution. The Fe-Al bimetallic particles are promising substances for further testing of Hg removal from gold processing waste solutions with different conditions and constituents. Keywords: Mercury removal, gold processing, Fe-Al bimetals, optimisation

Jan 1, 2020

By Y. G. Zhu, G. B. Zheng, X. X. Hu

Reverse flotation by cationic collectors is the classical method to reduce the silicon content in magnesite. But because of the low selectivity, there is a large loss of magnesite in the reverse flotation froth. To improve the efficiency of the collector, a novel collector named BK433 was developed in this paper. In order to determine the performance of the BK433 collector, a systematic experimental study was carried out with a high silicon magnesite in China. The suitable pH environment of slurry was confirmed, the depression effect of different depressants was investigated, and the flotation effect was compared with commonly used traditional collectors. The results show that the BK433 shows remarkable good selectivity and collection ability. After the 4 stages reverse flotation, the SiO2 content decreased from 3.76% to 0.25%, SiO2 impurity removal rate reached 94.69%. Meanwhile, the MgO content increased from 45.06% to 46.73%，the recovery of obtained concentrate (calculated by MgO) was 82.79%. Keywords: Magnesite, Desilicication, Reverse flotation, Collector

Jan 1, 2020

By C. R. Royal, B. E. Friedland

Energy & Densification Systems (EDS) has developed patented new technology crushing/milling equipment. The EDS Multishaft Mill is a compact vertical mill – utilising high-speed impacts to break down particles with large reduction ratios, low energy consumption and improved liberation. The milling action combines several processes to provide a multitude of output products to satisfy various specifications. Comminution often accounts for more than 60% of the total power consumption on a mine. Two of the key problems to be solved in comminution is to bring down the specific energy required to reduce particle sizes; and to improve liberation or improve the efficiency of the beneficiation and recovery of the required elements. The EDS Mill has shown extensive improvements in both these areas with energy efficiencies up to 75% better than conventional milling equipment and improving liberation beyond the standard beneficiation processes. Conventional comminution equipment requires massive, heavy parts to be moved or rotated, consuming a lot of energy in doing so and not inputting that energy into the breakage of particles. In the EDS Mill, ramping up the relatively small shafts to speed is all that is required. Once at speed, the momentum, together with the minimal amount of constant drive required, means that the energy can be transmitted directly and efficiently into breaking down the particles. The high-speed impacts allow for a shattering effect of the particles and this quickly generates a high ratio of fines. Furthermore, the single impact force breaks particles in a very different manner to conventional compressive force breakage. With compressive force breakage, the particle breaks from point to point. With high-speed impact force breakage, the force tends to migrate through the weakest areas of the particle, which tend to be the mineral boundaries. This, therefore, means that when the particles break, the various minerals are exposed and hence liberated. Keywords: Comminution, high-speed impact, liberation, energy efficiency, novel comminution

Jan 1, 2020

By B. S. Agbenuvor, R. Morrison, V. Jokovic, C. Antonio

Body breakage usually involves breakage of particles by compression between two surfaces or impacted against a rigid surface, with the latter of interest in this current study. This mode of breakage has been identified to be the most elementary size reduction process and the most effective breakage process in comminution devices. It has been found in the literature that the progenies produced from this form of breakage possess a residual velocity at which they move, after the breakage event. In the pursuit of efficiently utilising applied comminution energy, a laboratory-scale breakage device has been developed to deliver precise specific impact breakage using the concept of a fast-moving hammer and a slow-moving particle to effect breakage. The residual velocities of the progenies from the breakage event may be used to generate secondary breakage without further addition of energy if the particle encounters a stiff obstacle such as a steel plate at right angles to its path. To measure the extent of secondary breakage, the steel wall of the device was lined with a rubber material to suppress the second set of breakage events. The performance of the new device is analysed, based on calculated t10 values and measured residual velocities of the moving daughter particles. Testing involving anvils showed a much finer PSD and higher t10 compared to the test involving rubber lining. This finding is corroborated by the measured residual velocities of the progenies. The resulting energies calculated from the measured residual velocities of the progeny were found to be approximately 42-88% of the primary specific impact energy. This retained energy was found to be sufficient for secondary breakage. Keywords: Impact breakage, residual velocity, secondary breakage, energy utilisation

Jan 1, 2020

By D. Chetty

Process mineralogy forms an integral part of geometallurgy, with ore body characterisation being but one piece of the puzzle in optimising value derived from ore. This value is not only financial, but importantly speaks to sustainability from environmental and social standpoints. The reactive approach of traditional process mineralogy is increasingly being replaced by predictive methods in the geometallurgical sphere, for which mineralogical information must be timeously acquired on judiciously selected sample sets to understand ore variability and its impact on orebody processing. Such processes are not only metallurgical, but must incorporate mining and waste management protocols. Dwindling resources like water and energy also demand innovative approaches for lowering consumption and/or promoting recycling along the value chain. Addressing such challenges relies on appropriate characterisation of ore behaviour under different process regimes. Coupled with this, emerging characterisation tools are increasingly applied to understand the mineralogy of unconventionallypresented and difficult-to-process ores to meet societal and technology demands. Such tools are often accompanied by the generation of large, computationally intensive datasets, for which data management and storage are key requirements, particularly when deciphering temporal trends in ore and process variability. This paper will elaborate on these themes with examples that highlight the importance of process mineralogy throughout the minerals value chain. Keywords: Process mineralogy, ore characterisation, geometallurgy

Jan 1, 2020

By N. Mangadoddy, M. Padhi

The simulation of multi-component classification with modified viscosity model is the key focus in the current study. The classification of the ores using hydrocyclone is a renowned process and has a convoluted flow separation phenomenon. The highly turbulent flow inside hydrocyclone depends on the operating and design parameters apart from the feed mixture. Slurry viscosity also plays a dominant role in the classification of solid particles under the centrifugal settling action. A heterogeneous mixture with different components would result in different settling and rheological behaviour of particles’ suspension in the water medium. The newly proposed multi-component viscosity model (Mandakini et al. 2020) is implemented as a user-defined function to define the material property during the CFD study of the mixture rheology inside hydrocyclone classifier. The multiphase and turbulence model is solved using the modified Algebraic Slip Model and Large Eddy Simulation model. A comparison of the predicted component performance curves is evaluated with and without the corrected viscosity model. The density and size effect on the particle phase distribution inside the classifier are compared, where the misplacements of the finer particles are carefully assessed. Validation of the component classification in hydrocyclone with the laboratory experiments found a close agreement with the rheology based CFD data. Keywords: Dual-component, multi-component viscosity model, double-gap rheometer, computational fluid dynamics, high solids concentration

Jan 1, 2020

By S. Subramanian, L. Mweene

A siliceous low-grade copper ore from Mpanda Mineral Field (MMF), located in south west Tanzania consisting of SiO2 (58.3%), Fe (10.4 %) and Cu (0.9 %) was initially subjected to selective dispersion-flocculation using sodium trisilicate (STS) as a dispersant and xanthan gum (XG) as a flocculant. Subsequently, the selectively flocculated fraction was floated using a collector combination of potassium ethyl and amyl xanthate (1:1) in the presence of poly-ethylene oxide (PEO) as a silica depressant, to get a copper concentrate. The mineralogical investigations confirmed that the MMF ore was mainly composed of quartz, chalcopyrite and pyrite. Adsorption studies indicated that the adsorption density of XG on MMF ore decreased with an increase in pH, while that of PEO was independent of pH. The adsorption isotherms of XG and PEO on MMF ore exhibited Langmuirian behaviour. The zeta potential studies of MMF ore indicated that the iso-electric point (IEP) was located at pH 2.6, presumably due to its high silica content. In the combined presence of XG and STS, the zeta potential of MMF ore became more negative, with an increase in the concentration of STS. However, in the presence of PEO, the zeta potential of MMF ore became less negative with an increase in the concentration of PEO beyond the IEP, without shifting the IEP. Selective dispersion-flocculation of MMF ore improved the flocculated chalcopyrite grade to 19.8 %, yielding a degree of flocculation and a maximum separation efficiency of 12.1 % and 85.7 % respectively, after three desliming stages in the presence of 30 ppm of STS and 30 ppm of XG. A copper concentrate of 29 % grade and 90 % recovery containing silica of 0.9 % grade and 1.7 % recovery was achieved by flotation of the selectively flocculated component, using PEO as a depressant for quartz. Keywords: Flotation kinetics, flotation machines, flotation bubbles, particle size

Jan 1, 2020

By H. S. Baloyi, J. J. Rust

Spiral concentrators are widely applied to beneficiate commodities such as Cr2O3, mineral sands, and ferrochrome. The application of spirals for heavy minerals has been limited to a feed top size of 1 mm. At particle sizes larger than this, friction retards heavy particles and the particles' primary velocity approaches zero, resulting in a solids build-up, therefore inhibiting further separation. Recently, a new Multotec SC25 spiral model was introduced to the industry, which aids in beneficiating particles of up to 3 mm, with successes having been demonstrated in ferrochrome slag and titanium-bearing minerals. This spiral also has potential in electronic waste applications. This paper examines the applicability of the SC25 spiral in coarse Cr2O3 recovery. This stems from the fact that junior chrome miners tend to discard particles larger than 1 mm as they neither have milling nor dense media circuits. Also, alluvial chrome miners discard +1 mm even though this size fraction contains recoverable and liberated Cr2O3. The SC25 spiral has shown better Cr2O3 separation efficiencies than the SC21 model and with higher SiO2 rejection at saleable grades, thus providing an additional revenue stream specifically for the coarser (+820 μm) size fractions. Keywords: Spiral concentrators, chromite recovery, specific gravity, SC25 spiral, coarse spiral

Jan 1, 2020

By D. Mjonono, S. T. L. Harrison, N. Maliela, A. Kotsiopoulos

The improvement of coal waste management practices, through the co-disposal of waste rock with sulphur-lean fine waste tailings (<1 mm in size), can potentially mitigate acid rock drainage (ARD). On exposure to natural oxidants, the sulphide compounds present in mine discards generate acidity, creating a cyclical oxidation process intensified by iron and sulphur oxidising microorganisms. To inhibit the emergence of ARD, a stable mine dump may be constructed that is structurally stable and incompressible through the mixing of complementary waste fractions that reduce voids and access to acid generating surfaces. However, large voids manifest as scale increases, due to decreased particle interlock, enabling the unrestricted access of oxidants leading to the early onset of ARD. To avoid this phenomenon, enhanced packing approaches based on geochemical and geotechnical characterisation tests were employed to improve bed stability. These were assessed for long-term prevention using aggressive oxidative conditions typical of heap leaching applications. Favourable results were obtained for both blended and multi-layered systems. However, long-term success was most notable in beds with alternating layers of waste rock and fine waste tailings with co-mingled systems being more susceptible to failure. Deconstructed co-disposed beds provided valuable insight into post-leach structural transformations. Results indicated that relative changes in initial coarse discards to fine waste ratios were most notable in blended systems. Mass fractions of the comparatively more mobile fine wastes increased axially down these test columns that ultimately resulted in increased contact of oxidants with the acid generating waste discards. Conversely, more consistent ratios were observed in multi-layered configurations. In these systems, near impervious hardpans were created by the moisture retaining fine tailings, forming stable structures with neutralising barriers sustained at each alternating layer. Consistent with post-leach geotechnical characteristics, static acid-base accounting tests showed that the uppermost sections of blended systems were more susceptible to oxidation due to structural failure with fine particles migration, while the more stable multi-layered configuration evenly maintained neutralising capacity. Keywords: Acid rock drainage (ARD), bed structure, mobility of fine particles

Jan 1, 2020

By M. Asghari, R. Amiri, A. Haratian, S. Haji

The aim of this investigation is the evaluation of optimising the Shahroud copper ore beneficiation plant near the Iranian city of Shahroud, by means of a sensor-based ore sorter applying X-ray transmission technology. Currently, this plant’s feed capacity of 52t/h produces 0.5t/h copper concentrate with a grade of 30% and copper recovery of 85% from a feed grade of 0.36%. By applying an X-ray transmission-based (XRT) sorter, it is possible to preconcentrate the feed and remove the gangue materials before entering the plant, enabling the production rate of concentrate to be increased. In this study, the obtained representative sample from the plant was preconcentrated in particle size of - 75+10mm by using the X-ray sorter, thus the concentrate with Cu grade of 0.85% was obtained. Therefore, by using of X-ray sorter the Cu grade of plant feed was increased to 0.48%. By increasing the feed grade from 0.36% to 0.48%, and considering a Cu recovery of 85%, the production rate will be increased from 0.48t/h to 0.61t/h. Therefore, the preconcentration by using the XRT sorter increases the production rate by 27%. Keywords: Optimisation, preconcentration, copper, X-ray, sensor-based ore sorting

Jan 1, 2020

By L. M. Tavares, H. A. Petit, P. P. Cavalcanti

Iron ore pellets experience degradation during handling due low energy impacts and stresses. These impacts lead to volume or surface breakage of the pellets, creating fine particles that accumulate along the process. Some of the particles are in contact with the air and are dragged as dust into the workplace or into the atmosphere. The aerodynamic size of the particles allows them to be potentially inhaled by the workers or to be carried to nearby communities. Dust emission becomes critical at specific locations, such as transfer chutes and stockpiles. This work models the dust generation that occurs in an iron ore pellet stockyard using a methodology that combines experiments to quantify the degradation of the pellets to mathematical modelling of the dust emission mechanisms in the stockyard. First, laboratory tests were used to characterise the amount of mass and the particle size distribution of the dust generated by degradation under controlled conditions. Second, the flow over a stockpile of an iron ore pellet company was simulated with the aid of computational fluid dynamics (CFD) simulations. Third, semiempirical equations linked the data generated by simulations and laboratory tests to estimate the dust emission in the stockyard. Finally, results were validated with measured field data, obtained with a novel online measurement tool. Results suggest that the approach is able to predict dust emission with sufficient accuracy and can be used for further implementation in the management of operations and decision making in order to reduce the environmental impact of the process. Keywords: Iron ore pellets, abrasion, dust emission, mathematical modelling

Jan 1, 2020

By T. M. Louw, L. Auret, S. M. Bradshaw, F. F. Noelle

Modern industrial processes undergo frequent process state shifts due to changes in demand, feedstock or maintenance operations. Multimodal continuous process monitoring is therefore required for effective plant operation. Steady state analysis is a key component in effective multimodal monitoring. Various issues still remain in determining the starting and ending times of transition states. High dimensionality of process data is a characteristic that makes it difficult to select variables to consider when conducting steady state analysis. In order to address these problems, steady state detection on the principal component space is proposed. First, data undergoes principle component analysis to reduce dimensionality, then existing multivariate steady state detection algorithms are applied on the principle component scores. A novel algorithm, threshold PCA-SSD, is proposed and compared to an existing technique for steady state detection. The algorithms are investigated on high dimensional synthetic data Keywords: Multimode process, steady state detection, principle component analysis

Jan 1, 2020

By Yuandong Qiao, Erdong Li, Chao Lv, Junjian Ye

The gas-based CO reduction of Panzhihua vanadium titanomagnetite concentrate from 1100°C- 1250°C was investigated in this paper. The reduction degree increased with increasing temperature and increasing reduction time. At a temperature of 1250°C and a reduction time of 20 minutes, the metallisation degree reached 92%. The iron-titanium oxide phase transformation during reduction process was Fe2.75Ti0.25O5→Fe2.5Ti0.5O4→FeTiO3→FeTi2O5 according to the X-ray diffraction analysis. The SEM and map scan by energy spectrum of results of reduced samples at different temperatures indicated that the metallic iron particle size increased gradually as the reduction temperature increased. These results can serve as a reference for the reduction of vanadium titanomagnetite by CO. Keywords: Vanadium titanomagnetite concentrate, CO reduction, phase transformation, reaction mechanism

Jan 1, 2020

By D. T. Jordan, L. C. Coetzee, G. Kalenga, T. Mahlangu, W. J. Shipman

Although online free cyanide analysers are widely available and used in gold leaching operations worldwide, reliable automated free cyanide concentration control is not consistently achieved in many of these operations. Effective cyanide concentration control is complex owing to amongst others, the long residence times and ore feed variations. Model Predictive Control (MPC) has been proven to provide superior cyanide concentration control over traditional proportional-integral-derivative (PID) control, leading to savings in operational costs and reduced leachable gold losses. In the first example reported in this paper, MPC has improved control of the free cyanide concentration by keeping it within the required range 73% of the time during testing, compared to the previous PID controller that only achieved this target 30% of the time. This represents a 143% improvement. Another example demonstrates a 22% reduction in cyanide usage as a result of using MPC. MPC functionality is rarely available on the programmable logic controllers (PLCs) and control systems installed at gold plants, hence it is implemented on a personal computer (PC) networked to the cyanide analyser and plant infrastructure. Mintek has developed an embedded MPC that installs into the Mintek Cynoprobe analyser, adding MPC functionality without the need for additional PC and network infrastructure, reducing installation costs. This allows plants to minimise their expenditure on additional infrastructure while gaining the benefits of superior MPC technology for multiple cyanide dosage points. This paper will present results showing that the embedded MPC performs as well as Mintek’s existing PC-based StarCS MPC controller, making it a viable alternative for infrastructure-constrained gold plants. Keywords: Model predictive control, gold cyanidation

Jan 1, 2020

By K. C. Corin, M. S. Manono

The drive to reduce water usage globally has led to the mining industry reassessing its water usage - owing to past environmental degradation due to effluent discharge, tailings disposal and process water seepage into the water-table (Haggard et al., 2015). Factors such as water scarcity and environmental restrictions have created the necessity for recycling of process water in mining and other industries. Water recycling, however, causes an accumulation of ions in the process water, which may affect the strength of reagent action on the target mineral and in turn critical aspects of flotation may be impacted. Prior studies have shown that certain ions have specific effects on flotation: for example, calcium and copper have been found to enhance the adsorption of depressant onto gangue (Feng et al., 2013, Wang et al., 2016). Calcium and sulfate ions have been seen to give a higher concentrate grade than other ions present in process water (Manono et al., 2012). An increase in ion concentration will increase the mineral grade due to a decrease in entrainment (Corin et al., 2011). This study considers the impact of ion accumulation on reagent action and flotation response. Keywords: Flotation chemistry, water quality, ionic strength

Jan 1, 2020

By G. Mucsi, F. Kristály, R. Szabó

Kaolin is one of the most important ingredients in porcelain manufacturing. Its main phase is clay mineral containing aluminum hydrosilicate, namely kaolinite. Beside kaolinite it also contains quartz, illite and feldspars, which impair its applicability. Kaolin is a non-reactive material, but metakaolin can be produced by heat treatment (calcination) of kaolin, which is also used by the cement and concrete industry due to its latent hydraulic (pozzolanic) properties. Beside thermal activation, reactive materials can also be produced by mechanochemical activation in general. This paper deals with the mechanochemical activation of kaolin. The raw material was classified into various size fractions, then ground in a batch stirred media mill under varying circumstances (residence time, circumferential velocity). After grinding, the structural changes of the products were investigated by X-ray diffractometry (XRD) and Fourier Transform Infrared Spectroscopy (FTIR). In addition, the particle size distribution of the ground material was determined by laser scattering. As a result of the grinding experiments and analytical measurements of the products, it can be clearly seen that reactivity of kaolin can be improved significantly by mechanical activation in a stirred media mill, but due to its high energy density the circumstances were optimised carefully to avoid aggregation. Keywords: Kaolin, grinding, stirred media mill, mechanical activation, material structure

Jan 1, 2020

By B. T. McKenna, H. Nirschl, P. Morsch, B. Fränkle

Generally, filter media used in filter presses have a limited operational life. Besides mechanical damage, there is a growing deposition of fine particles with increasing filtration cycles, the so-called blinding, inside the filter cloth. The resulting decreasing pore diameters not only affect the efficiency of the filtration process itself by increasing the filter medium resistance, but also make it more difficult to regenerate the filter press between the individual cycles, e.g. leading to an insufficient cake discharge. Since tailings filtration requires handling of large process flows containing a significant fraction of particles in the lower micrometer range, an investigation into media rejuvenation is of interest. Keywords: Tailings filtration, filter fabric resistance, filter media regeneration, filter cleaning, intraweave contamination

Jan 1, 2020

By N. Naudé, J. de Villiers, C. P. Bergmann

It is proposed that the physical and chemical composition of individual mineral particles are an appropriate basis from which to model the performance of minerals processing equipment. Such a particle-based approach enables modelling across disciplines such as comminution, classification and flotation and automatically generates the traditional mass and chemistry balances. The general approach followed is to generate full grade-recovery curves for a separator by sampling product and tailings at varying mass yields. AutoSEM characterisation of each sample provides particle data that is used to reconcile mineral deportment and liberation to the measured size and assay data. In this way, partition surfaces can be generated and modelled. In this paper, an empirical model for the performance of a gravity spiral concentrator treating fine chromite ore is proposed. The model is based on the combination of two partition surfaces and is able to accommodate both forward and reverse gravity concentration effects per size fraction. The results show that this approach is well suited to represent spiral performance for the cases evaluated. The effect of two feed particle size distributions on the performance of the Multotec SC21 and SC25 spirals is demonstrated. Simulations are run in an Excel environment, linked to Python-enabled routines for enhanced data management and calculation speed. Keywords: Spiral concentrator, gravity separation, particle-based modelling

Jan 1, 2020

By M. A. Belokrys, F. O. Milovich, D. D. Aksenova, V. A. Ignatkina, V. A. Bocharov, A. A. Kayumov

Copper remains one of the most sought-after metals in the industry. Tennantite (Cu12As4S13) and tetrahedrite (Cu12Sb4S13) that occur in pyrite copper, copper-zinc and polymetallic ores are an additional reserve for increasing technological parameters of flotation. Pyrite mass (85% and more) has similar flotation properties to other sulfide minerals, and imperfection of reagent regime is a common problem that occurs during quality copper concentrate processing. Due to tennantite's and chalcopyrite's differences in flotation properties with butyl xanthate at pH > 10, tennantite often reports to the flotation tails. Hight alkalinity of the medium increases the oxidation level of sulfide minerals, thus increasing the concentration of S2O3 2-, SO3 2- SO4 2-. At processing facilities in the Russian Federation, the concentration of thiosulfate ions in pulp is in the range of 0.5 to 1.5 g/l and higher, which leads to a decrease in flotation activity of the tennantite. The effect of butyl xanthate (as standard), thionocarbamate, dithiophosphate and M-TF (a mixture of dithiophosphate and thionocarbamate) on monomineral (tennantite and pyrite) flotation in the presence of sodium thiosulfate was studied. The variable parameters of the "thiosulfate-collector" flotation complex were thiosulfate concentration (100-900); pH (8-12,5). The scientific methods included contact angle for hydration assessment; TEM for surface studies; UV spectroscopy and iodometric titration for solubility control. Comprehensive studies have established a connection between the concentration of thiosulfate ions and tennantite flotation selectivity. Flotation activity of tennantite is higher at pH 8 with a thiosulfate concentration under 300 mg/l. Pyrite remains active with butyl xanthate under the same conditions. TEM images of the surface films of the collectors on pyrite were acquired at pH 8. The MTF provides the biggest difference in adsorption, wettability and flotation activity between tennantite and pyrite at pH 8 and thiosulfate concentration under 300 mg/l, in comparison with butyl xanthate. The results of the research will be applied to improve selective reagent regimes of tennantite minerals flotation. Keywords: Tennantite, pyrite, chalcopyrite, sphalerite, flotation, sulfhydryl collectors, thiosulfate, adsorption, mechanism, reagent regime

Jan 1, 2020

By F. Wang, P. Rosario, A. Vazirizadeh

Advances in data science, machine learning, and artificial intelligence are transforming mining and mineral processing, and making them more algorithm intensive. The paradigm is shifting from one of detection and control to one of prediction and optimisation. To build up reliable prediction and optimisation models these steps should be passed: collect, aggregate, cleanse, and process large amounts of complex, structured, and unstructured data as a first stage. Next, the right tools must be used to extract meaningful features from the operational data for the build-up of a prediction model. A grinding circuit is one of the challenging areas in terms of determining its overall efficiency and the prediction of product quality. In a processing plant, some of the KPIs are typically unmeasurable or infrequently measured for real-time application, e.g. ore grindability and grind size. However, these variables are required to calculate the grinding circuit efficiency and to model the operation performance. In this paper, advanced analytic tools have been used to extract one selected KPI (cyclone overall P80 or final grind size) and its relationship with measured variables for an industrial dataset. The results illustrate that the performance of the grinding circuit can be predicted for real-time monitoring by means of measured data. Keywords: Grinding circuit, data-driven, machine learning, process modelling, hybrid modelling

Jan 1, 2020