Birchtree is an ultramafic ore body and prior to 2003, represented one third of the production for Inco's Thompson operation. The ultramafic host rocks are high in MgO which, when processed through the smelter, promotes problems in the roasters and converters due to the high liquidus temperature. In October 2003, production from Birchtree mine ramped up to more than half of the total feed to the mill. The mill cannot produce a suitable concentrate for the smelter with current circuit configuration and operating practices. The target for the mill was to produce a nickel concentrate with less than 3% MgO. / Continuous mini flotation cells were used to explore processing options. Two campaigns were carried out. The first looked at rejecting the magnesium silicate minerals with the use of reagents from the rougher and scavenger concentrates. Reagents included sodium silicate, guar gum and two types of CMC. Results indicated that current plant practice of CMC addition to the scavenger circuit favours minimizing MgO grade in concentrate. / The second campaign focused on rejecting MgO minerals from the rougher-cleaner concentrate with the use of a dilute CMC solution. Flotation retention time, CMC dosage rate, and conditioning time, were varied. Results indicate that 0.1 g CMC per kilogram of solid in slurry produced a concentrate grading <3.0% MgO.

Overbreak and damage to rock walls is one of the most serious problems encountered in blasting operations. Several techniques have been developed to control the undesirable effects of rock blasting. These techniques are collectively known as wall-control blasting methods. The stress distribution around pressurized holes has been numerically evaluated, in order to analyze the mechanism of wall-control blasting methods. The effect of blast geometry and the role of discontinuity on this stress field has also been studied in detail. The results obtained by numerical modelling have been verified by controlled blasting experiments, and further supported by analysis of existing roadcuts on a large scale. [...] / Durant les opérations de dynamitage, les problèmes rencontrés les plus sérieux som les dommages causés aux parois et le bris de profil. Plusieurs techniqutes ont été développées pour contrôler les effets indésirables du dynamitage des roches. Ces méthodes sont connues sous le terme “méthodes de sautage pour le contrôle de parois”. Une évaluation numérique de la distribution des contraintes autour de trous sous pressurion a été réalisée, dans le but d’analyser le mécanisme du sautage purle contrôle de parois. Une étude détaillée a été entreprise sur l’effet de la géométrie du dynamitage ainsi que le rôle de la discontinuité sur le champ de contraintes. Les résultats obtenus par modélisation numérique ont été vérifiés avec l’aide d’essais de dynamitage contrôlé et confirmés par une analyse à grande échelle des coupes de roc existants le leng des chemins. [...]

A 7.6 cm (3") Knelson concentrator was used to estimate free gold content in samples extracted from industrial grinding and gravity circuits. Its recovery was comparable with that of amalgamation and the Mozley Laboratory Separator (MLS). It was then used for a study on gold grinding and classification behavior. / Laboratory studies of monosized gold and silica showed that gold produces fewer fines upon grinding: 75% of the mass reports to the next Tyler class, as opposed to 45% for silica. Disappearance from the monosized class (840-1200 $ mu$m) follows first order kinetics for both minerals. However, the rate constant of gold was five to six times lower than that of silica, and folding predominated over actual breakage. Folding yields either spherical or cylindrical particles which flatten upon additional impacting. / Smearing of gold onto silica and embedding of silica into gold flakes were observed. It was postulated that gold smearing can lead to metallurgical losses in flotation circuits. A gravity circuit could then increase overall gold recovery. / The Knelson concentrator was used to study the behavior of gold in the Hemlo mill grinding circuit. The ratio of the selection function of ore to gold increased from 6 at 50-100 $ mu$m to 20 at 500-1000 $ mu$m. In the primary cyclones, gold's cut size was much finer than the ore, 20 vs. 57 $ mu$m.

One of the key factors in ensuring safe slopes in open pit mines is the control of the groundwater flow within the slope. When analyzing the flow regime and its characteristics, traditional numerical methods designed for soil-like materials may not always apply. In the present thesis, the two-dimensional finite element code FlowD was developed to analyze steady-state seepage through fractured rock masses, under saturated conditions. The program provides the user with two modelling options: porous (soil) materials, or fractured rock masses. The double-porosity model was incorporated in the code, in order to better model flow through anisotropic, heterogeneous fractured rock masses, where Dirichlet and Neumann boundary conditions can apply. FlowD program calculates total pressure head, flow gradients and velocities within the specified domain. Drains, drainage galleries and wells, as well as aquifer recharge can be simulated. A real large-scale case study, with complex geological features has been simulated in order to demonstrate the application of the double porosity model. Three different scenarios have been modeled for the same slope, which are natural groundwater regime, vertical well simulation, and drainage galleries simulation. The results show good agreement with the predictions of both the consultant and the mine's engineering division.

This thesis deals with the flow mechanics of hydraulic and high density mine backfill slurries in pipelines. Various empirical, rheological and mechanistic approaches are presented and analysed along with some aspects pertinent to experimental testing of backfill slurries. / The main contribution of this work is, particularly, in the development of an analytical model to describe the flow and predict the pressure gradient of a class of high density backfill whose motion in pipelines follows the Plug Flow Model (PFM). The development of the model called for investigating the conditions required for establishing Plug Flow. It was found that mix proportioning procedures, similar to those found in the concrete industry, are key factors in obtaining Plug Flow. / Pressure drop was found to be a function of the thickness of the Bingham plastic annular layer surrounding the cylindrical core of aggregates. Analytical equations were proposed to solve for the thickness of this layer by considering the rheology of the mixture. Alternatively, the thickness of the annular layer may be estimated by considering the relative proportions of the mixture with respect to aggregates void content. The model offered pressure drop predictions in good agreement with published data. The proposed model may also serve as an alternative to Mooney's method, when dealing with the annular lubricating layer effect characterising mixtures in Plug Flow.

This study examines the issue of economic optimization associated with mine development and extraction decisions, focusing on the choice of an appropriate installed capacity as well as the control of the cutoff grade and capacity utilization variables. It shows how capacity and production strategies lead to an enhanced economic outcome of resource development and extraction. / An important aspect of the study is the development of deterministic algorithms for cutoff grade and production rate optimization within existing mine facilities. These variables are determined by maximizing the present value of future benefits associated with mine production. Two opposing economic forces are at play in the optimizing process. A higher present value can be achieved by increasing operating profits through a higher cutoff grade and/or extraction rate. This strategy decreases the losses due to discounting. However, a higher cutoff grade generally leads to higher operating costs, and may also result in the loss of ore. A higher production rate within an existing mining system, likewise, increases unit operating costs. These reduce operating profits and act to lower the present value. / Both static and dynamic optimization methods are developed. Static optimization refers to determining the constant cutoff grade and production rate that maximize the total discounted operating profits. Dynamic optimization refers to the situation in which periodical adjustments are allowed for the production variables. In this case, declining cutoff grade and production rate schedules within an established mining system yield higher total discounted profits over the static solution. In general, the schedule is controlled by the discount rate, as well as by trade-offs between ore quality and reserves, and between production rate, cutoff grade and production costs. / The study subsequently focuses on the choice of installed capacity at the mine development stage. The objective here is to find the installed capacity level that maximizes the net present value of the project. The consideration of capital costs alone tends to limit the increase in installed capacity. The preproduction period has a similar effect. When both variables are combined, the increasing preproduction period and capital costs associated with higher levels of installed capacity impose a severe burden on the maximization process, causing limitations to the rising trend of optimum installed capacity at high discount rates.

Ground preparation design encompasses problems of ground characterization, equipment selection and performance. Ground characterization has in the past been the subject of empirical diggability classification schemes based on case studies. A new system is proposed for evaluating the ease of excavation of different ground conditions, is put forward. This has been applied successfully in different field studies. While such empirical schemes are useful to mining design, there still remains a basic lack of understanding of machine-ground interaction in the excavation process. This will be essential for long term machine and method advancement. / This thesis reviews the approach and methodology adopted for the development of predictive digging models in mining. Past theoretical attempts to model soil-rock cutting processes are critically evaluated and an analytical digging model applicable to fragmented rock is proposed. Physical modelling studies, concentrating on the interaction between machine and spoil characteristics, are reported. These considered in particular the digging profile propagated by cable shovel operation, the impact of geotechnical parameters on the profile, and the efficiency of digging in fragmented rock. These were undertaken in a controlled laboratory environment based upon an instrumented scale model and photographic studies. The salient parameters influencing the efficiency of the digging processes are identified and a bucket filling model for cable shovels is proposed.

In mineral processing, the liberation distribution of the mineral phases in an ore is an important factor. Unfortunately, liberation distribution is very difficult to measure accurately. One problem is the stereological effect associated with a microscopic liberation analysis. / Many methods have been proposed to correct for the stereological effect, but they remain essentially untested, in part due to the lack of a standard or reference material. In this work, a method was established to create an artificial standard material. / Silica grains were placed in an epoxy resin to create a block of material that was crushed to yield liberated and locked particles. The composition, and thus liberation, of these particles was determined with a series of heavy liquid separations. / By changing the grain size to particle size ratio, the type and amount of locking was affected. The best compromise between the amount of locked material produced and the production of simple locked particles (which pose the greatest stereological challenge) was found; it occurred at the point where the grain and particle sizes were the same. / Liberation analyses were performed on this material and compared with a model prediction based on the sectioning of spheres with single planar interfaces. There were some discrepancies between the analysis and model results.

Tumbling mills describe a class of mechanical systems defined by a cylindrical chamber filled with balls and/or rocks that rotate around their own longitudinal axis. Due to the action of charges, the lifters/liners wear significantly and their shape will be changed which will influence the efficiency of the mill. Liner bolts are the mill components which connect the mill interior with the exterior environment. With this property, bolts have been paid more and more attention. The first part of the thesis covers the structural analysis of the lifter/liner bolts, and the possibility for on-line charge motion measurement with the aid of bolts is discussed by comparing the strain distributions at the bolt areas where the bolts occupy different mill circumferential positions. In order to measure the lifter/liner wear, a real-time, on-line wear sensor was developed by the Comminution Dynamics Laboratory. However, the positioning of the sensor/sensors affects the strength of the mill is unknown. The second part of the project provides detailed research on this issue by FEA analysis so as to provide a basis for the practical application of this kind of sensors. Finally, during the grinding process of tumbling mills, some mills can vibrate greatly at some ranges of mill rotating speeds and as a result the mills cannot work properly because of this vibration or resonance. With six different mill models, the final part of the thesis investigates the mill natural frequencies and their modes by FEA and furthermore, explores the effect of mill diameter, length on these frequencies and modes.

Since blasting is a comminution process, the feasibility of establishing a correlation between the Bond rod mill work index (kWh/t) and some easily measurable physico-mechanical properties relevant to blasting was investigated. Further, the concept of operating blast work index and its potential applications were explored. Four different rock types and a well documented case study of two blasts were selected for this study. / The work index is found to be uncorrelated with the density and unconfined compressive strength, slightly related with tensile strength, and well correlated with dynamic rock properties, especially the P-wave velocity and the bulk modulus. The standard deviation in measured compressive and tensile strength values is found to be very high, compared to their seismic and dynamic elastic properties. / The case study encompassed two blasts with identical blast-patterns and rock type consuming almost the same amount of explosive (kg/t) but with very different specific blast energies (MJ/t). The agreement between operating work index of the two blasts (13.4 kWh/t vs. 13.1 kWh/t) and laboratory work index (17.0 kWh/t) was modest (within 30%). The operating work index corresponding to either blast has been used to predict the product size (P$ sb{80}$) of the other. The choice of feed size (F$ sb{80}$) was discussed; whereas the previous researchers had used the theoretical value of infinity, the use of much smaller value, the average of effective burden and spacing, was proposed. This concept is used in a proposed method to estimate blasting energy requirements as a function of blast geometry, rock type and desired product size. Future test work that would provide a data base and validation for this concept is described.