A t10-based method for evaluation of ore pre-weakening and energy reduction

Currently the effect of the pre-weakening of ore particles by high voltage pulses is evaluated by the percentage change of A 1 b values between pulse-treated and untreated ore particles. The values of A 1 b, widely used as an ore breakage competence indicator in the mineral industry, are determined from the parameters of the JKMRC breakage models. In this study a t₁₀-based method was developed to predict the degree of size reduction, t₁₀, of pulse-treated particles from that of untreated particles broken at the same size/energy level. This method incorporates one parameter, C_Ab, which is equivalent to the percentage change of A 1 b values.The t₁₀-based method was validated using nine sets of comparative JK Rotary Breakage Tester data on pulse-treated and untreated ore samples over a wide range of impact specific energies and particle sizes. The t₁₀-based method can be used to calculate the energy reduction due to the pre-weakening effect in the downstream comminution process. It indicates that the energy reduction by pre-weakening increases with an increase in the target product fineness and the degree of pre-weakening, and with the decrease in feed particle size.     

Ore impact breakage characterisation using mixed particles in wide size range

The Drop Weight Tester (DWT) for ore impact breakage characterisation uses particles in five narrow size fractions, and the JKMRC Rotary Breakage Tester (JKRBT) uses four size fractions, both with three impact energy levels for each size fraction. It is time consuming to prepare these narrowly sized particles and to carry out size analysis on the 15 DWT or 12 JKRBT products, so a Wide-size JKRBT characterisation method was developed. In this method, the mixed particles in 13.2–45 mm size range are tested as one size class in the JKRBT by single-particle breakage mode. The wide-size feed is then divided into several virtual narrow size fractions by simulation, based on which the impact product size distributions are calculated using a size-dependent breakage model. Four sets of measurement data, consisting of two feed samples in the 13.2–45 mm size range with different size distributions tested with two impact energy levels, are adequate to determine the three model parameters. In the case where a benchmark ore of known breakage characteristic parameters is available, one Wide-size JKRBT impact treatment can determine the ore competence change parameter using a t₁₀-based model.     

Characterisation of superficial breakage using multi-size pilot mills

Low energy surface breakage has a high frequency of occurrence and thus plays a significant role in grinding processes. Yet this superficial breakage is poorly understood, measured and modelled – forming the focus of this work.Pilot mills of 0.8–1.8 m diameter, designed to provide a predominantly surface breakage environment with efficient removal of the resultant progeny, are utilised to characterise superficial breakage. A new rate, that of superficial breakage (1/(kW h/m²)), is introduced which measures fractional superficial breakage rate per energy provided to the surface of the material. This methodology is proposed as being suitable for understanding and characterising the surface breakage behaviour of ores.Tests were conducted on two ores with different hardness. Superficial breakage rates varied from 2 to 16 (1/(kW h/m²)) for the different ores and mill sizes, indicating a good sensitivity to ore type and the need to understand the applied stress – related to mill size. The results show that a single ‘surface breakage rate for use in mill modelling is incorrect as the rate of superficial breakage is dependent on the size of the mill and therefore the inter-particle stressing conditions.     

Modeling breakage rates in mills with impact energy spectra and ultra fast load cell data

A new era in modeling particle size distribution in grinding mills started at the beginning of 2000s. A direct estimation of breakage parameters became possible via computation of collision energy by discrete element method (DEM) and material breakage data.The material breakage data can be obtained for primary modes of breakage. In this study, impact and abrasion are assumed to be the primary modes of particle breakage, which are readily studied in the laboratory. The impact breakage mode is studied in a drop-weight apparatus and in a specialized device known as the ultra fast load cell. The abrasion mode of breakage is studied in a laboratory scale ball mill. Next, the particle breakage versus energy data is converted into breakage rates via impact energy spectra of the grinding mill computed by a DEM code. The fundamental material breakage information is converted into energy based breakage distribution function.The verification of the modeling concepts is shown for a 90 cm laboratory scale ball mill. In the batch mill, approximately a 10 kg mass of limestone in the 30 mm size is ground with around 100 kg of 50 mm steel ball charge. The breakage rate and the breakage distribution functions constitute the parameters of the energy based batch population balance model. It is shown that accurate particle size distribution predictions are possible with this modeling approach for different grinding regimes.     

Effects of media shape on milling kinetics

Spherical balls are the dominant grinding media used in ball mills. However, balls which are initially spherical, wear into non-spherical fragments. The proportion of worn, non-spherical balls in the charge of a mill fed with 50 mm balls is dominant in ball sizes less than 30 mm. Their effects on mill performance in terms of material breakage are not yet established.The variations of specific rate of breakage with single size feed and fractional filling U, were studied for the two media shapes (spherical and worn balls). Higher breakage rates were noted with spherical media than worn balls but the differences narrow with decreasing feed size and increasing material fractional filling, U.     

The effect of ball size on breakage rate parameter in a pilot scale ball mill

The objective of this study is to investigate the effect of ball size on grinding kinetics in a pilot scale ball mill. Six different ball media gradings were tested. Comparative tests were conducted in batch ball mill having 1.2 m diameter and 0.6 m length at constant operating condition of mill such as media mass, mill speed and input specific energy. Feed samples were ground batchwise and representative sample was taken from inside the mill for each determined grinding period. Grinding process in ball mill was modeled and the specific rate of breakage was calculated for the each test. The results indicated that the relationship between different breakage rate and particle size has a maximum for each ball size distribution. Consequently, a new equation to correlate maximum ball size and particle size at maximum breakage occurs is proposed.     

Towards optimising ball-milling capacity: Effect of lifter design

The effect of liner/lifter profile on kinetics of batch grinding and the milling capacity in general was assessed using mono-size quartz material of 30 × 40 mesh (−600 + 425 μm) as feed. The liner profiles tested were, (i) bevel with 60° lifter face angle to represent the new liners, (ii) bevel with 45° lifter face angle to represent the worn liners and (iii) worn bevel modified with cone-lifters. The tests were conducted under identical conditions to allow a comparative analysis of the results. In all cases, the breakage followed the first order hypothesis. The experimental size distribution data was well predicted by the S and B model, thus allowing for estimation of breakage and selection parameters (i.e., γ and a_T) for the three liner situations tested. The optimised values of the specific rate of breakage, S_i for the three liner profiles tested were 0.381, 0.287 and 0.365 min⁻¹, respectively, which clearly indicates the benefit of cone-lifters. The breakage distribution function (B_ij) values did not vary significantly with liner profile, which echoes the findings by other researchers.     

Effects of slurry filling and mill speed on the net power draw of a tumbling ball mill

The pool of slurry is known to lower the power drawn to the mill. An attempt to ascertain this observation by relating load orientation to mill power for a range of speeds and slurry fillings was undertaken.To this end, a Platinum ore (−850 μm) was used to prepare a slurry at 65% solids concentration by mass. The Wits pilot mill (552 × 400 mm), initially loaded with 10 mm balls at 20% volumetric filling, was run at 5 different speeds between 65% and 85% of critical. The net power draw and media charge position were measured. After this, the slurried ore was gradually added to the media charge for slurry filling U between 0 and 3. A proximity probe and a conductivity sensor mounted on the mill shell provided a means of measuring both the position of the media charge and that of slurry. The data collected for the load behaviour and net power draw was later analysed.It was found that Morrell’s model could not fully explain the effect of slurry volume on net power draw especially for an under-filled media charge (i.e., for U < 1). The size of lifters and grinding balls used could be the reason for this. That is why a piece-wise function was curve-fitted to the power data to help make sense of the inconsistencies observed.     

Use of the attainable region method to simulate a full-scale ball mill with a realistic transport model

Optimisation and better control of milling circuits require extensive modelling of milling data. This paper extends the enquiry to the use of the attainable region (AR) technique to determine the optimal residence time of ore in a ball mill. It also evaluates the energy requirements of the mill at the set residence time to maximise the production of the desired size range which enables maximum recovery of platinum group minerals (PGM) during the flotation stage.With these purposes in mind, the breakage function and the scaled-up selection function parameters were used to simulate the operating conditions required by an industrial ball mill and the power requirements were predicted using the Morrell power model. This allowed the application of the AR methodology to be extended to a full-scale ball mill. Then a link was established between residence time to mill product specifications for a given feed size.The findings showed that the residence time required by a full-scale mill falls between those at which the fully mixed and the plug flow mills operate. The results also showed that operating the ball mill at a lower mill speed and a higher ball filling saves energy. Mill speed was again found to be a key operational factor for controlling the retention time of particles inside the mill. This yielded valuable insight for the importance of optimally controlling both the residence time of the material inside the mill and the amount of energy required to maximise the desired size range, in this case −75 + 9 μm.     

An analysis of the energy split for grinding coal/calcite mixture in a ball-and-race mill

Interactions among components in the heterogeneous grinding would change energy consumed characteristics of components if compared with those in the single-component breakage. In this paper, energy split phenomenon for the coarse grinding of super clean anthracite coal (SCAC)/calcite mixture of 2.8–2 mm in the ball-and-race mill is investigated. Before the analysis of experimental results, accuracy of energy split function in terms of time-dependent breakage rate is first discussed. Energy consumed characteristics of grinding in the ball mill and ball-and-race mill are also compared. Breakage model of product t₁₀ (yield of progenies in −0.237 mm) vs specific energy is used to describe the energy-size reduction of the single-component and multi-component grinding. Interaction between components is reflected by the comparison of specific energy of components in mixture and single breakage to yield the same product t₁₀. Based on the energy balance, energy split factors of components in different time and mixed conditions are first determined. This parameter shows no change with time. Calcite increases the grinding efficiency of SCAC significantly, with the energy split factor for SCAC ranging from 0.68 to 0.73, which means less specific energy is consumed by SCAC to yield the same t₁₀ if compared with the single breakage. As the volumetric ratio of calcite increases in mixture, grinding energy efficiency decreases and energy split factor of calcite increases from 1.70 to 1.83. Soft material reduces the grinding energy efficiency of hard one in the multi-component breakage.     

Analysis of grinding media effect on specific breakage rate function of particles in a full-scale open circuit three-compartment cement ball mill

A full-scale three-compartment FLSmidth® cement grinding ball mill with dimensions of Ø3.5 × L10 operating in open circuit was sampled to analyse the grinding media effect on specific breakage rate function of particles. Size reduction performance of the ball mill was evaluated with respect to the applied grinding media size. Samples from the circuit and inside the mill were collected. Mass balance of the circuit was done using JKSimMet Steady State Mineral Processing Simulator. Specific discharge and breakage rate functions of particles were estimated using perfect mixing modeling approach (Whiten, 1972) on the basis of the proposed open circuit three-compartment ball mill model structure (Genç and Benzer, 2015). Maximum specific breakage rate was related to maximum grinding media size in the grinding compartments. An exponential correlation was found to exist between maximum grinding media size and maximum specific breakage rate. Relationship between maximum grinding media size and maximum particle size was also fitted to an exponential function. Findings indicated that, grinding performance of cylpebs applied in the third compartment did not improved the size reduction performance as compared to the grinding performance of the first and second compartment.     

A study of atmospheric acid leaching of a South African nickel laterite

Nickel and cobalt acid leaching from a low-grade South African saprolitic laterite using sulphuric acid was studied. Ore characterisation was performed by XRD and XRF. Batch agitation leaching tests were conducted at atmospheric pressure investigating main parameters: particle size and percent solids at 25 °C and 90 °C. Ore characterisation showed that the ore is a saprolitic laterite with nickel present in lizardite. Leaching tests showed that nickel and cobalt could be leached from the ore at atmospheric pressure. Nickel was found to be more leachable from the coarser −106 + 75 μm fraction, with 98% Ni being extracted at 90 °C after 480 min. Cobalt was not favoured by variation in particle size and increased percent solids. Increasing ore percent solids improved nickel extraction at 25 °C however at 90 °C extraction decreased due to a diffusion layer build-up as a result of amorphous colloidal silica. The co-dissolution of magnesium and iron was elucidated. Nickel leaching data at increased temperature and percent solids fit the shrinking core model equation, k_dt = 1−2/3x − (1 − x)^2/3 showing that nickel leaching reaction was diffusion controlled under the set conditions.     

Hydrodynamic and kinetic characterization of industrial columns in rougher circuit

In the present investigation the relationship between collection zone rate constant (k_c) and gas dispersion parameters, viz. bubble size (d_b), superficial gas velocity (J_g), gas hold-up (ε_g) and bubble surface area flux (S_b) was evaluated. Experiments were conducted in an industrial (4 m in diameter and 12 m high) and a pilot (0.1 m in diameter and 4 m high) flotation column in rougher circuit at Miduk copper concentrator in Iran. Gas hold-up was measured using pressure difference technique and mean bubble sizes were estimated from a drift flux analysis. It was found that the collection zone rate constant was not correlated with d_b and J_g solely but was linearly dependent on ε_g and S_b for the range of interest. Collection efficiency (E_k) and floatability factor (P) in the industrial columns were quantified (E_k = 3.1%; P = 7.7 × 10^?3). The influence of operating parameters comprising superficial gas velocity, slurry solids% and frother dosage/type on S_b and flotation kinetics was discussed. Analysis of available industrial data suggested that S_b and ε_g were related by S_b = 4.46ε_g over the range 30 < S_b < 60 s^?1 and 7% < ε_g < 14%.     

Dissolution kinetics of primary chalcopyrite ore in hypochlorite solution

Dissolution kinetics of primary chalcopyrite ore from Artvin-Murgul, Turkey has been studied in hypochlorite solution. The dissolution kinetics was found to be controlled by diffusion through the product layer as the rate-determining step. The activation energy (E) was calculated as 19.88 kJ mol⁻¹.     

The effects of chamber diameter and stirrer design on dry horizontal stirred mill performance

This paper focussed on investigating the effects of chamber diameter and stirrer design on cement grinding performance of a horizontal type dry stirred mill. Within the scope, pilot scale test works were undertaken with two different chamber diameters (20.4 cm and 26.4 cm) having the same length and three different stirrer designs (wing, cross and disc) having the same diameter (16 cm). The chamber diameter tests were performed at the same stirrer design, media size and media filling. The studies concluded that, the use of larger chamber improved the grinding efficiency since 31.8% and 35.8% less energy was consumed than the smaller mill at the RR_d50 of 1.41 and 1.66 respectively. This behaviour of the larger mill can be attributed to the increased gap distance between the chamber wall and stirrer edge. With regards to stirrer design, the statistical evaluations, grinding results and temperature measurements all indicated that the disc design of stirrer ground the particles more effectively at high energy levels (>40 kW h/t). The use of the disc design reduced the energy consumption by 21% (at RR_d50 of 3.5). This was attributed to dissipation of energy as heat since the temperature measured for the wing and cross types were higher than the disc type.     

Comparison of energy efficiency between ball mills and stirred mills in coarse grinding

Stirred mills are primarily used for fine and ultra-fine grinding. They dominate these grinding applications because greater stress intensity can be delivered in stirred mills and they can achieve better energy efficiency than ball mills in fine and ultra-fine grinding. Investigations were conducted on whether the greater performance of stirred mills over ball mills in fine grinding can be extended to coarse grinding applications. Four different laboratory ball mills and stirred mills have been tested to grind seven ore samples with feed sizes ranging from 3.35 mm to 150 μm. A case study on full scale operations of a 2.6 MW IsaMill replacing the existing 4 MW regrind ball mill at Kumtor Gold Mine in Kyrgyzstan is also included. This paper summarizes the major findings from these investigations.     

Determination of the milling parameters of a platinum group minerals ore to optimize product size distribution for flotation purposes

Most concentrators desire to operate under optimal design configuration that guarantees high mineral recovery and low operational costs. The optimal design configurations are determined through studying the material to be milled in a laboratory mill under standard conditions. This is achieved through determining the selection and breakage function parameters and applying the mathematical simulation of the grinding process in order to optimize the size reduction process. The desired particle size is determined by the downstream processes, in our case, flotation. To this end, three mono-size classes feeds 850–600 μm, 600–425 μm and 425–300 μm of a platinum ore were ground using three different ball sizes (10, 20 and 30 mm) in a laboratory mill for the grinding times 0.5, 1, 2, 4, 8, 15 and 30 min. The data collected was used to determine breakage and some of the selection function parameters. The remaining parameters were back-calculated within the population balance model framework. The parameters were then used to obtain the product size distribution (PSD) that was later compared with the experimentally measured one. The milling kinetics for the desired size class for flotation was also simulated.There was a good match between the predicted and the experimentally measured PSD. The results of the milling done for further 60, 90, 120 and 240 min to validate the simulated milling kinetics from the determined parameters also showed good match between the simulated and the experimental one. This further confirms the validity of the determined parameters. From this, it becomes possible to determine the grinding conditions for optimal flotation.     

Froth zone characterization of an industrial flotation column in rougher circuit

In this investigation the froth zone of an industrial column (4 m “diameter” × 12 m “height”) in rougher circuit was characterized. Experiments were carried out at Miduk copper concentrator, Iran. Miduk is a unique copper processing plant which utilizes columns in rougher circuit. Cleaning and selectivity actions in the rougher froth were illustrated using solids and grade profiles along with RTD data. The impact of froth depth (FD) on overall rate constant (k) and k–S_b relationship was evaluated. Dependency of overall flotation kinetics on froth depth and gas velocity (J_g) was modeled by k = 4.97(FD)^?0.87(J_g)^0.80. Froth recovery (R_f) was estimated and modeled in terms of froth residence time of slurry (FRT_Slurry) as R_f = R_f,maxexp(?k × FRT_Slurry). Finally, the correlation between k, S_b (indicative of the collection zone performance) and FRT_Slurry (indicative of the froth zone performance) was modeled by k = 0.02 (FRT_Slurry)^?0.62(S_b)^0.82.     

Boundary conditions for gas rate and bubble size at the pulp–froth interface in flotation equipment

This paper describes the effective boundary conditions for the gas dispersion parameters of bubble size, superficial gas velocity and bubble surface area flux, in mechanical and column flotation cells. Using a number of previously derived correlations, with appropriate simplifying assumptions, and experimental data reported from plant practices, the boundary conditions were identified. Thus, it was shown that these constraints typically allow for a mean bubble diameter range of d_b = 1–1.5 mm and superficial gas rate of J_g = 1–2 cm/s, in order to maximize the bubble surface area flux, S_b = 50–100 s⁻¹. Under these conditions there is no carrying capacity limitation, while keeping a distinctive pulp–froth interface.