Tumbling Mill Research Articles

Transformation of the seeded aluminum hydroxide into sinter-active α-alumina

Mechanical treatment of aluminum hydroxide containing 0.5-5wt.% α-Al2O3 nanoparticles has been carried out in a planetary ball mill at various milling intensities. It was shown that, when the ball acceleration developing in the mill is equal to or exceeds 10 g, the subsequent calcination at 800°C is sufficient for the complete transformation of the hydroxide into α-alumina with a crystallite size of less than 100 nm. After calcination of the ball-milled powder in the flow of humid air, the alumina particles were somewhat larger, had a spherical shape and did not form strong agglomerates. A body obtained by dry compaction of the powder calcined in the flow of humid air sinters up to 98% of the theoretical density at 1350°C. Nearly the same results were obtained when a simple tumbling mill was used, although a much longer duration of mechanical treatment was needed..

Feature extraction based on the fractional Fourier transform for vibration signals with application to measuring the load of a tumbling mill

Feature extraction plays a major role in data preprocessing and soft sensing. In this paper, a feature extraction method based on the fractional Fourier transform (FrFT) is proposed to estimate the load of a tubular ball mill. The FrFT, a generalised form of FFT, can draw intermediate time–frequency representations for a signal. Vibration signals are first transformed into a frequency spectrum by using the FrFT with multiple-orders, yielding multi-order frequency spectrums (MFS). Then, according to the sensitivity thresholds under different wear conditions, multi-order characteristic frequencies (MCF) are selected from the MFS, and some statistical characteristics are extracted from MCF as initial features. Only informative features can significantly benefit the forecasting accuracy. Therefore, some candidate features used to build the soft-sensing model are chosen from statistical characteristics. Finally, the load of a tumbling ball mill can be obtained by fusing the results of multi-order regression models based on an adaptive-weighting coefficient. Compared with some feature extraction methods and the performance of the proposed method in the Fourier domain, experimental results show that the feature extraction method proposed is effective in modelling the load parameters of a tubular ball mill.

Effect of Chemical Additives on Reduction in Mill Power During Continuous Grinding of Chromite Overburden Materials in a Tumbling Mill: A Case Study

Tumbling mills account major share of power consumption in mineral processing industries. In this paper, an attempt is made to minimize the power consumption in continuous grinding of chromite overburden by improving the grinding performance with different chemical additives. The additives taken are calcium chloride, sodium hexametaphosphate, sodium hydroxide and sodium silicate. It is found that sodium silicate is most effective in reduction in power consumption. The enrichment of nickel in sample is studied with grinding time using the same additives. The rheology of mill is also observed with different additives. In all observations, sodium silicate is found to be most suitable among four additives used.

Preliminary validation of a new way to model physical interactions between pulp, charge and mill structure in tumbling mills

Modelling of wet grinding in tumbling mills is an interesting challenge. A key factor is that the pulp fluid and its simultaneous interactions with both the charge and the mill structure have to be handled in a computationally efficient way. In this work, the pulp fluid is modelled with a Lagrange based method based on the particle finite element method (PFEM) that gives the opportunity to model free surface flow. This method gives robustness and stability to the fluid model and is efficient as it gives possibility to use larger time steps. The PFEM solver can be coupled to other solvers as in this case both the finite element method (FEM) solver for the mill structure and the DEM solver for the ball charge. The combined PFEM-DEM-FEM model presented here can predict charge motion and responses from the mill structure, as well as the pulp liquid flow and pressure. All cases presented here are numerically modelled and validated against experimentally measured driving torque signatures from an instrumented small-scale batch ball mill equipped with a torque meter and charge movements captured from high-speed video. Numerical results are in good agreement with experimental torque measurements and the PFEM solver also improves on efficiency and robustness for solving charge movements in wet tumbling mill systems.

Evolution of a generic, dynamic and multicomponent tumbling mill model structure incorporating a wide-range 4D appearance function

This work aims to build a generic dynamic model structure, which can accommodate interchangeable sub-models of each sub-process, making it amenable to continuous upgrade without the need for redevelopment, for multicomponent tumbling grinding mills. The Generic Tumbling Mill Model Structure (GTMMS) is based on a population mass balance framework which incorporates breakage characteristics, transport, classification along the mill, a discharge function, and energy consumption incorporated in a dynamic mill model structure. Version III builds on two earlier versions by incorporating energy distributions derived from discrete element modelling, an updated version of the 4D breakage appearance function which applies to a broader size range, and addresses multi-component ore breakage via the probability distribution of energy split based on material stiffness. The model has been tested against multi-component plant survey data. GTMMS III suggests a mechanistic insight into mixture prediction through component analysis and is a step forward towards the unified comminution model (UCM) with its mechanistic, generic, and dynamic prediction capability.

Effect of different grinding media composition on the flotation of a PGM ore

In the present study, the effect of using different grinding media, viz, ceramic (CE), forged steel (FS), 15% Cr steel (CS), stainless steel (SS) and a mixture of stainless steel and 15% Cr steel balls (50:50) (SS + CS) were investigated using a Platinum Group Metal (PGM) ore, UG2. Test work was conducted using a laboratory scale horizontal tumbling mill and a batch flotation cell. Power draw to achieve a specific P80 was determined for each media type. The different flotation behaviour obtained from different media is discussed in terms of particle size distribution, water recovery, solid recovery and valuable mineral recovery. pH, Redox potential (Eh) and dissolved oxygen (DO) level were also measured. Grinding products of five different media produced the same solids and water recovery, but the grades and recoveries of PGM and chromite were different. Grinding products of ceramic media had the best flotation performance, which produced the highest PGM recovery and the lowest Cr2O3 recovery. Meanwhile grinding products of stainless steel balls (SS) produced the lowest PGM recovery and highest Cr2O3 grade and recovery in the concentrates. The paper proposes possible reasons to explain these observations.

Experimental Study of Charge Motion in a Tumbling Ball Mill

Tumbling ball mills are a common comminution device in the mineral industry processing, wherein the particle size reduction is performed by action of the grinding media. Different forms of transverse motion in a rotating cylinder have been studied by many researchers. The aim of this paper was to study the transitional phenomena between cascading, cataracting and centrifuging motion. In order to observe these transitions, experiments were carried out in a tumbling mill of 300 mm diameter, and three sizes of chrome steel balls were used as grinding media. Each size of ball was employed in three different lengths of the mill with the purpose to investigate the end-walls effect in the behavior ball charge. The image analysis technique was employed to identify the transition regions. A high-speed camera with maximum speed of 2000 frames/s was used to record images. Experimental data showed it was not found the end-wall effects on the transitional behavior, and this effect can be assessed in numerical studies by analyze of the specularity coefficient.

Investigating grinding media differences in microstructure, hardness, abrasion and fracture toughness

It is recognised that grinding media wear can represent up to 50% of the operating costs of a given tumbling mill. Over the years, a number of works have explored the development of different ways and means to both understand grinding media wear as well as model and predict it. The focus of the present work is to examine the differences in microstructure, hardness, abrasion and impact toughness of grinding media from eight different manufacturing sources. Results will be presented for 125mm diameter media typically used for SAG mills. A discussion will address any issues highlighted by the results than may contribute to predictive wear model development as well as indicate possible directions for future research.

Comparative study of mechanical activation of molybdenite (MoS2) with and without magnesium (Mg) addition

Molybdenite concentrate was mechanically activated in the presence of magnesium metal using a planetary mill in order to try to form elemental molybdenum. A sample milled for 90 min in planetary mill showed almost the same structural distortion as after 50 h in a tumbling mill, neither showed the presence of molybdenum metal. However, molybdenite milled together with magnesium showed more amorphization than separately milled molybdenite. Despite the apparent lack of reaction, molybdenite showed higher solubility in dilute HCl after milling with magnesium than without.Mongolian Journal of Chemistry 16 (42), 2015, 30-33

Development of a novel methodology to determine mill power draw

This paper presents an approach that has been developed to calculate the power draw of each size within a distribution of charge in a tumbling mill. Positron Emission Particle Tracking (PEPT) and the Discrete Element Method (DEM) are used to test the methodology. Experiments are conducted using dry spherical glass bead charge in a laboratory scale tumbling mill, which is mounted with a torque transducer and tachometer to provide measurements of mill power. Particle tracking information from PEPT is used to reconstruct the motion of glass beads and infer the overall charge behaviour, while DEM is employed to simulate particle motion and interaction, with collision mechanics calculated using the Hertz–Mindlin contact model. For both sets of data, the product of torque and average angular velocities in discrete cells are accumulated to obtain mill power. This method is found to be within statistical agreement with measured power for all cases investigated.

Coupled DEM-CFD Model to Predict the Tumbling Mill Dynamics

The charge motion in a tumbling mill has been mostly described by empirical, mechanistic and computational models. The computational model presented in this work is a three phase approach for the tumbling mill that combines a particle description for the solids modelled by the Discrete Element Method, and the continuum description for the fluid by CFD approach. In the present work a phase coupled approach is developed using C++ subroutines to model the charge and slurry dynamics inside a tumbling mill by mapping the particles on the CFD mesh and resolving the particle volume and velocities on per cell basis. The coupled DEM-CFD approach is implemented and the effect of drag force on the slurry by the particle motion. The set of coupled simulation were run varying the slurry rheology and results were validated with equivalent PEPT experiment of lab scale mills and a very good agreement is found in some cases. The Beeststra drag correlation was used to calculate the drag force between the fluid and the particles. The free surface profile of the charge as well as the slurry is calculated as well as the axial center of mass profile of the mill.

How to account for operating condition variability when predicting liner operating life with DEM – A case study

Wear prediction is important in the development and optimisation of liner design, and for helping to manage reline strategies. DEM is able to predict wear and profile evolution of liners over the life cycle for specific sets of conditions. However, real operation conditions often vary – usually in ways that are not known in advance, which limits the ability to use such model predictions in managing associated wear. In this study, we use a specific case of a tumbling mill to explore issues around the variability of wear caused by different operating conditions within the applicable range for the mill, and consider how these can be included in a modelling framework in order to assist with mill liner management.

Development of a method to measure charges in tumbling mills

Tumbling mills are often taken as the object of optimization studies because they are a type of equipment that consume large amounts of energy. Among the current available resources to conduct such studies, mathematic modelling presents great efficiency due to its low cost, speed and reliability. The total charge and grinding media charge are very important variables to conduct modelling exercises that aim at power draw and product size distribution forecasting. However, the common measurement methods require people entering the equipment, which carries a number of adversities related to confined spaces. In this regard, this paper presents the development of a method and the prototype of a device able to measure tumbling mill charges, quickly, precisely, with low cost and, above all, ensuring safety. The result of this work is a method that allows equivalent or superior precision in comparison to the existing methods, whose main aspect is to eliminate the requirement of people entering dangerous environments, such as tumbling mills.

A novel method for modelling of interactions between pulp, charge and mill structure in tumbling mills

Modelling the pulp fluid and its interaction with both the charge and the mill structure is an interesting challenge. The interaction is normally modelled with a combination of CFD and DEM, where the DEM particles (grinding balls) create the structure through which the fluid penetrates, and in its turn creates forces on the grinding balls. However, in a tumbling mill, many free surfaces are found and that limits the use of CFD. An alternative computational approach is here necessary.The smoothed particle hydrodynamic (SPH) method has earlier been used to model a ball charge and its interaction with the mill structure. In the present contribution, a SPH description of the pulp fluid is introduced. The lifters and the lining are still modelled with the finite element method (FEM), and the grinding balls with DEM. This combined computational model makes it possible to predict pressure within the pulp fluid. It is also possible to predict how the dampening effect of the pulp liquid is affected by its viscosity and density. The charge induced torque in a laboratory-scale ball mill is used for validation, and the mechanical shock waves travelling in the mill system are described.

Validation of tumbling mill charge-induced torque as predicted by simulations

Understanding mill charge motion is important. In the charge, the center of gravity is shifted from the rotational center of the mill system, and its motion is induced by rotation of the mill. At the same time, the charge creates a torque in the mill system. Breakage of ore particles and wear of liners/ball media are closely linked to this motion. To study these phenomena in a physically correct manner, numerical models for different parts of the mill system are needed. Validations of such models are scarce, because of the difficulty of measuring inside a tumbling mill. Experimental measurements in a lab mill were done for a number of load cases: varying feed material, mill filling, mill speed and pulp liquid. The mill is set up to measure the charge-induced torque. The accuracy is good, with relative uncertainty less than ± 2% for relevant load cases. A full three-dimensional numerical model of the whole mill is used to predict induced torque. Agreement between predicted and measured torque at steady-state is good. In addition, the model can accurately predict the mill startup behavior for torque and mill power. This proves that the model is physically correct, and can be used for modeling large-scale mills.

A Neuro-fuzzy Model with Clustering Initialization to Estimate the Tumbling Mill Efficiency

A Neuro-fuzzy Model with Clustering Initialization to Estimate the Tumbling Mill Efficiency

DEM-PBM approach to predicting particle size distribution in tumbling mills

Tumbling mills using balls as grinding media are used extensively in the mining and cement industries to produce fine powders; however, it is well known that the process of size reduction is highly energy-intensive. For this reason, much of today’s comminution research is aimed at understanding grinding mechanisms, by estimating power draft and modeling milling circuits using the discrete element method (DEM). DEM predicts individual particle trajectories, the distribution of energy and contact forces at each collision, lifter wear and power draw. This paper presents a prediction of the discharge particle size distribution using three components, mass balance model (sometimes referred to as a population balance model, PBM), impact energy distribution of the mill obtained from the simulation of the charge motion using DEM and breakage characteristics of particles determined from drop-ball tests DEM simulation of the charge of a tumbling mill was done using the code MinProSim 1.0 to obtain impact spectra at different operative conditions. Then, thepopidation balance model (PBM), based on impact energy distribution, was used to predict the final particle size distribution in a batch mill and compared with experimental data. It is demonstrated that the DEM-PBM approach allows a reliable solution to predict particle size distribution for tumbling mills. Ideally, the model has to predict size distribution for any operational condition and feed characteristic; however, it has been shown that predictions include some degree of error.

Comparisons of PEPT derived charge features in wet milling environments with a friction-adjusted DEM model

The extent to which the discrete element method (DEM) represents particulate flow in wet tumbling mill environments is investigated by adjusting the friction coefficient in the simulation to best match positron emission particle tracking (PEPT) experiments of the same media in steady-state, re-circulating fluids of water and slurry. Porosity and velocity distributions of 5mm, mono-sized glass spheres in a 300mm diameter mill are statistically compared to an equivalent dry DEM model. These show a very high degree of spatial agreement with RMS differences of only 0.11%. Typically, around 80% of the mill volume is found to have identical results. Similar levels of agreement are found for a range of operating speeds from 31% to 75% of critical speed. These demonstrate that a solid particle DEM model of the charge, without an explicit representation of the slurry, is able to accurately capture the motion of the solid charge. Agreement is improved by calibration of the wet friction coefficient. The DEM results indicate that the flow is not particularly sensitive to the exact choice. Small differences between DEM and PEPT are observed in the free surface part of the flow. The presence of fluid slightly slows the fall and compaction of the charge leading to a more slowly moving and higher free surface in the middle of the mill. The flow is also found to be highly three dimensional with significant axial flow generated by the end walls of the mill.

Load Behavior Prediction in a Tumbling Mill

mportance of the SAG (Semi Autogenous Grinding) mills for the mineral complexes persuades many researchers to study the behavior of these huge equipments. Laboratory mills and the DEM (Discrete element method) software are the prevalent and reliable methods of modeling the mill behaviors. The former is very costly and the latter not only takes the much time but also needs the very advanced computers. In the present work, a method is presented to predict the charge profile and shoulder and toe positions for a case study mill. Experimental relations are manipulated to determine the en masse geometry and the kinetic relations are used to explain the ball movement on the liner in order to determine the shoulder and toe positions. Model results have very good accordance with the experimental results obtained of a laboratory mill. The model is used to predict the load behavior of a SAG mill in the Sarcheshmeh copper complex in Iran. Acceptable results due to low cost and quickness make it as an appropriate method for predicting the charge profile in that tumbling mill.

Kinetics of leaching zinc and indium from indium-bearing zinc ferrite mechanically activated by tumbling mill

Synthetic indium-bearing zinc ferrite (IBZF) was mechanically activated by a tumbling mill. Subsequently, the behavior and kinetics of leaching zinc and indium from the initial IBZF and activated IBZF in sulfuric acid solution were investigated. The results demonstrate that the tumbling ball milling enhances the leaching reactivity of IBZF and increases the zinc and indium extractions. The reaction temperature and H2SO4 concentration have an important effect on leaching zinc and indium from the initial IBZF and activated IBZF. The kinetic model -ln(1-x) = kt was used to describe the kinetics of leaching zinc and indium from the initial IBZF and activated IBZF and to calculate the kinetic parameters. The tumbling ball milling causes a decrease in activation energy and reaction order, which indicates that the dissolution of IBZF after mechanical activation by a tumbling mill becomes less sensitive to temperature and H2SO4 concentration. For the zinc leaching, the activation energy decreases from 73.9 to 70.3 kJ/mol and the reaction order decreases from 0.72 to 0.70. For indium leaching, the activation energy decreases from 78.1 to 73.5 kJ/mol and the reaction order decreases from 0.82 to 0.69. It is found that the dissolution rates of the initial IBZF and activated IBZF are controlled by surface chemical reactions.