The main binding force in wet iron ore green pellets has been found to be the cohesive force of the viscous binder. The wet compression strength (wet-CS) in green pellets is, however, also influenced by the green pellet plasticity. A certain degree of plasticity is needed to sustain the green pellet growth rate. Too much plasticity results in decreased bed permeability and production problems. As the plasticity increases, wet-CS decreases. The amount of moisture needed to create a given degree of plasticity depends on particle properties and on the particle size distribution. Therefore, it was of interest to study how wet-CS would be influenced by variations in raw material fineness, if the green pellet plasticity was kept constant, i.e. the green pellet properties would be compared under relevant industrial balling conditions. For this purpose, magnetite concentrates of different particle size distributions were balled in a laboratory drum and the moisture content for constant plasticity was determined for each of the materials. No difference in green pellet wet-CS as a function of the raw material fineness was found when the bentonite binder was used and the plasticity was adjusted to a constant level. Green pellets prepared of raw materials with narrow size distributions were just as strong as those with broader ones. This is because the main binding force is the cohesive force of the viscous binder. In green pellets balled without the bentonite binder, wet-CS increased with increasing specific surface area in the raw material, in a similar manner as has been shown in earlier agglomeration literature. In this case, the capillary forces prevail. Comparison of wet-CS at constant moisture, instead of constant plasticity, would lead to erroneous conclusions. Fineness, or rather the slope of the particle size distribution curve, had a major impact on the moisture content needed for constant plasticity. If the slope increases, more water is needed to keep the plasticity on a constant level. Implications of these results in control of industrial iron ore balling circuits are discussed.

Thermal volume changes and oxidation mechanisms in magnetite iron ore green pellets balled with 0.5% bentonite binder, as a function of raw material fineness and pellet porosity, are shown. When a pellet starts to oxidize, a shell of hematite is formed around the pellet while the core still is magnetite. Dilatation curves were measured under non-oxidizing and oxidizing atmospheres to separately describe thermal volume changes in these two phases. Dilatation measurements showed contraction during oxidation between 330 and 900 °C by 0.5%. The extent of contraction was not influenced by the raw material fineness or the original porosity in pellets. Sintering started earlier in the magnetite phase (950 °C) compared to the hematite phase (1100 °C). The sintering rate increased with increasing fineness in the magnetite concentrate. A finer grind in the raw material would, therefore, promote the formation of duplex structures with a more heavily sintered core pulling away from the less sintered outer shell. At constant porosity in green pellets, the oxidation time became longer as the magnetite concentrate became finer, because of the enhanced sintering. In practical balling, however, the increase in fineness would necessitate the use of more water in balling, which results in an increase in green pellet porosity. These two opposite effects levelled out and the oxidation time became constant when green pellets were balled at constant plasticity. Combining the results from the oxidation and dilatation studies revealed new information on the rate limiting factors in oxidation of iron ore pellets. At 1100 °C, the diffusion rate of oxygen was limited by sintering in the magnetite core, taking place before oxidation rather than by the diffusion rate of oxygen through the oxidized hematite shell, as has been claimed in earlier literature. The oxidation rate was at maximum at around 1100 °C. At 1200 °C, the rate of oxidation substantially decreased because both the hematite shell and the magnetite core show heavy sintering at this temperature. Dilatometer measurements showed large thermal volume changes in the presence of olivine, at temperatures above 1200 °C. This is explained by the dissociation of hematite back to magnetite. Dissociation leads to an increase in the volume of the oxidized shell, while sintering of the magnetite core is further enhanced by the olivine additive.

The properties of iron ore green pellets with varying additions of a surface-active flotation collector reagent (Atrac) were studied by small-scale balling. The compression strength and plasticity were measured with a semi-automatic measuring device and the pressure curves were saved and subjected to further mathematical treatment. The green pellet breakage was also filmed with a high-speed camera. Adding Atrac to the pellet feed seriously damaged the quality of green pellets, even in small dosages. This is because an increasing amount of air bubbles became so strongly attached on the particle surfaces that they could not be removed during compaction by balling. The adsorption of air in green pellets was seen as an increase in porosity and a decrease in the filling degree (proportion of pores filled with water). Both the wet and dry compression strength decreased. The air bubbles behaved in wet green pellets like large, plastic particles and the plasticity increased beyond an acceptable level. Breakage started inside the green pellets, along the air bubbles, and generated multi-breakage patterns in wet as well as dry green pellets. Green pellet breakage to crumbs instead of a few distinct segments, promotes the generation of dust and fines and leads to lower bed permeability in the pelletizing machine. The results show that the decrease in iron ore green pellet wet strength in the presence of surface-active agents is not fully described by the so called Rumpf equation, where surface tension and contact angle are used as variables to describe the capillary forces. The green pellet breakage in the presence of air bubbles took place by crack propagation along pore structures rather than through the loss of the capillary forces.

Fundamental research during the past decade has been focussed on understanding the role of viscous forces on agglomerate deformability and strength. Much of this work has been done on glass spheres using Newtonian liquids as a binder. In this work, we show the variations in plasticity and strength of magnetite iron ore green pellets with varying liquid saturations and binder dosages (viscosities). For this purpose, a new measuring instrument was built to analyze the green pellet wet compression strength, plastic deformation and breakage pattern. Industrial iron ore green pellets are over-saturated and a supporting "network" of viscous liquid is formed on the green pellet surface. At least half, probably more, of the total binding force appeared to be due to the cohesive force of the network. The other half (or less) of the total compression strength can be explained by the capillary force. Due to irregularities on green pellet surfaces, both fully developed concave pore openings and saturated areas are expected to be found at the same time. Wet green pellets started showing plastic behaviour as they became over-saturated. In over-saturated green pellets, an explosive increase in plasticity with increasing moisture content was seen, due to the contemporary increase in porosity. Plasticity is an important green pellet property in balling and should gain the status of a standard method in green pellet characterization. It is suggested that the control strategy for the balling circuits be based on plastic deformation and compression strength of green pellets instead of the rather inaccurate drop number. The results also point out the importance of knowing whether the balling process should be controlled by adjusting the moisture content (plasticity) or by adjusting the bentonite dosage (viscosity). These two operations are not interchangeable-even if they would compensate in growth rate, the green pellet properties would differ. A new green pellet growth mechanism is suggested, based on the measured over-saturation. Firstly, green pellet plasticity needs to exceed a minimum level to enable growth. This limiting plasticity defines the material-specific moisture content needed in balling. Secondly, it is suggested that the growth rate be controlled by the viscosity of the superficial water layer rather than by the mobility of the pore water.

In this work, the purpose was to study the impact of variations in feed ore properties on the performance of a primary autogenous grinding circuit by ore characterisation and simulation. Samples were selected to represent various points in the production system; ore faces with different drillability, grinding circuit feed, mill charges and waste rock. The investigation was carried out at the LKAB Kiruna mine in northern Sweden.The result clearly shows that self-breakage occurs ahead of the mill since the ore hardness, or resistance to breakage, increase with the distance from the mining face. Ore from a location, which by the mine is characterised as “hard to drill”, has the lowest resistance to breakage, and the surrounding rock is clearly harder than the magnetite ore. Validation of a simulation model for the primary autogenous grinding circuit reveals that the differences between simulated and experimental data are small. Therefore, the model is used to simulate the influence of variations in feed ore characteristics on the circuit performance. The simulations show that the net throughput from the circuit at a coarse–hard feed will be ≈10% higher compared to a situation when the feed is fine–soft. Moreover, a fine and soft feed results in a coarser particle size distribution of the mill discharge, compared to a coarse and hard material. However, it is the amount of coarse material in the feed, which is the most influential factor.

A model-assisted method is used for designing the control strategy for the new LKAB hematite plant in Malmberget (Sweden). Steady-state and dynamic process simulations successively help to select strategic controlled and manipulated variables, and specify the corresponding instrumentation. The control system implemented in the plant reflects the conclusions of the study. It includes an optimizing control system based on specified software, with a fuzzy logic expert system, a dynamic model of the process and an optimizer.

The effect of chemical variables on the kinetic parameters of apatite flotation from magnetite has been investigated. The two common first-order kinetic models reported in the literature, i.e., the model with rectangular distribution of floatabilities and the model with fast and slow-floating components (F-S model) have been applied in the evaluation of flotation results. The models are evaluated by fitting the flotation results from batch flotation tests. The results although indicate that both models describe the apatite flotation kinetics well in a wide range of kinetic parameters, the F-S model is found to be better in the goodness of fit to the results from every flotation test and also to describe the flotation performance adequately. The kinetic parameters of the F-S model are varied with a change in chemical variables. In particular, the rate constants ratio Kf/Ks is found to be an important parameter for achieving selectivity between apatite and magnetite, and the effect of reagents dosages on the ratio is discussed. The effect of particle size on the apatite flotation kinetics is also illustrated by analysing the froth products at each flotation time in different size classes.

Dephosphorization of magnetite fines by flotation in such a way as to minimize the fatty acid coating left on the magnetite surfaces is the principal technical challenge facing the mineral processing division of LKAB, Sweden. Modelling of flotation kinetics enables the influence of chemical and operational variables on apatite flotation from magnetite fines to be predicted. Five typical first-order flotation kinetic models are evaluated by statistical techniques, after an estimation of model parameters by a nonlinear least-squares fitting program, with use of the authors' own results and results taken from the literature. Evaluation of the models is based on two aspects: goodness of fit to the experimental results and suitability to describe the flotation behaviour. The first-order models with a rectangular distribution of floatabilities and with fast and slow floating components (F-S model) gave an excellent fit to the experimental results of apatite flotation when compared with the other models, the latter being superior not only in the goodness of fit to apatite flotation and other mineral flotation results but also in its description of flotation behaviour. The kinetic parameters (percentage recoveries and rate constants of fast and slow floating particles) in the F-S model have a physical significance and can thus be used for qualitative as well as quantitative interpretation of flotation performance. The influence of collector dose can be effectively included in the F-S model by making realistic simplifications and relating the collector dose to the fraction of slow floating particles. The ratio of fast and slow rate constants can be used to describe the selectivity of apatite flotation from magnetite.