The present work attempts to determine the feasibility of the accelerometer as an end-point control sensor for the basic oxygen steelmaking process. In this study, two series of laboratory scale BOF simulations were performed. In the first series, the accelerometer was sampled at low frequency to obtain an amplitude-time relation that can be related to the rate of decarburization during the oxygen blow. In contrast, the accelerometer was sampled at high frequency in the second series in order to discern the presence of specific vibrational frequencies that can be related to the process. As a final aspect of the research work, several high frequency simulations were terminated prematurely in an attempt to elucidate the carbon content of the bath at the point in time when the accelerometer can clearly detect the final change in the rate of decarburization.

An investigation of copper matte vacuum purification was undertaken. A mathematical model using monatomic, diatomic sulphide and diatomic oxide vapours of the impurities was developed to explain the vacuum refining process. The model was used to predict the overall refining rates for variables such as temperature, matte grade, oxygen activity and chamber pressure of the melt. / A series of experiments was undertaken to characterize the dust produced during vacuum refining of copper matte. An attempt to selectively condense the vapours produced during the vacuum refining experiments was carried out. / The "Lift-Spray" vacuum refining process was used to refine 15 to 40 kg batches of molten copper matte containing 35 to 78% copper. The removal rates of lead, bismuth, arsenic, antimony, selenium, nickel and silver were measured under different levels of matte grade, chamber pressure, lifting gas flow rate and dissolved oxygen content in the melt. The ranges of the variables under study were from 10 to 600 pascals chamber pressure, 0 to 40 normal milliliter per second of lifting gas flow rate and 10$ sp{-16}$ to 10$ sp{-7}$ atmosphere of oxygen activity. / In conclusion, LSV refining of copper matte was shown to be a promising process. Furthermore, scale-up to industrial size is now possible. The scale-up dimensions compare well to the dimensions of the RH degassing unit presently used in the steel industry. (Abstract shortened by UMI.)

With the aim of simulating the hot rolling process, single and double hit compression tests were performed in the temperature range 800 to 1200$ sp circ$C at strain rates of 0.2 to 50 s$ sp{-1}$ on selected low carbon steels containing niobium, boron and copper. The stress/strain curves determined at high strain rates were corrected for adiabatic heating and constitutive equations were formulated. When dynamic recovery is the only softening mechanism, these involve a rate equation, consisting of a hyperbolic sine law, and an evolution equation with one internal variable. When dynamic recrystallization takes place, the incorporation of the dynamically recrystallized fraction in the above evolution equation makes it possible to predict the flow stress after the peak. / The kinetics of static recrystallization were characterized in terms of the mean flow stresses, which lead to more accurate results than alternative methods. Appropriate expressions were formulated for the recrystallization kinetics and the mean austenite grain size as a function of predeformation, temperature and particularly strain rate. / Particular attention was paid to prediction of the temperature distribution through the thickness of the rolled plate or strip. The effects taken into account are radiation and convection from the surface when the material is between stands, and conduction to the rolls and the temperature increase due to mechanical work when the material is in the roll gap. An explicit finite difference method was used to calculate the temperature distribution through the thickness of the workpiece during processing. / On the basis of the temperature model and the constitutive and recrystallization kinetics equations, a computer model for the prediction of multi-stage rolling force and microstructural evolution was developed. The predictions of these models are in good agreement with measurements on both experimental and commercial steels. Also, by appropriate control of the thermomechanical parameters, high strength steels with transition temperatures below $-$100$ sp circ$C were developed. These results constitute a step towards the on-line control of plate and hot strip mills, and the development of new tough high strength steels.

Solid residues are generated at various stages of metal production. Some of these residues still contain significant quantities of metals. In determining possible process routes for residues a detailed knowledge of the mineralogy (i.e. the minerals or phases present) is required. This thesis examines the mineralogical characteristics of four residues: copper smelter dust, jarosite residue and hot-acid-leach residue all from the metallurgical plant of Kidd Creek Mines Ltd., and low-acid-leach residue from the Canadian Electrolytic Zinc plant. / The methodology of characterization included the fractionation of the residues based on particle size, magnetic susceptibility and leachability in sulphuric acid. The fractions were analyzed by X-ray diffractometry and electron beam techniques. / A major phase in all the residues was zinc ferrite. The characterization of this phase showed that it does not have the stoichiometric composition of ZnFe$ sb2$O$ sb4$ but exhibits a variable composition of the type Zn$ sb{x}$Fe$ sb{3-x}$O$ sb4$. This variable composition corresponds to a measured variation in magnitude susceptibility of the ferrites. / Some treatment routes for these residues are examined; special emphasis was placed on the potential application of magnetic separation. The hot-acid-leach residue appeared the most amenable to magnetic separation.

In order to design the metal delivery system for the single belt caster in the MMPC (McGill Metals Processing Centre) foundry, water modeling and mathematical modeling were carried out for a newly devised three-chamber type tundish. Water flow in the acrylic tundish was visualized using dye injection. Flow velocities were also measured using a Dual Nd-YAG PIV (Particle Image Velocimetry) system. A commercial FEM code, FEMLAB 2.3 was adopted to predict the velocity field and temperature profile within the tundish, especially in the output chamber. Calculated results were validated with the PIV measurements. A full-scale water model was built for the single belt caster to simulate the casting operation and to validate the optimized delivery system. Temperature profiles for the tundish wall were also predicted to choose a suitable method of preheating and to determine refractory wall specifications. / A three-chamber type tundish comprising an entry chamber, a head control chamber and an output chamber was designed to provide clean metal and strips of uniform thickness across the width of the belt. An output chamber proved to be essential for removing the bubbles and for obtaining a uniform film of water on the substrate by preventing strong hydraulic jump. The output chamber had to be completely closed for rapid bubble removal. For rapid filling of the output chamber at start up, the starting stopper proved to be essential. The 3-hole type nozzle, proved to be more effective for removing the bubbles, was found to have problems in terms of strongly impinging jet flow and non-uniform lateral velocities. / Using mathematical modeling and full scale water modeling, including PIV measurements, the "FD" type nozzle, which had a multi channel flow modifier in the output chamber and a slot type inlet nozzle, was found to be the best in terms of rapid bubble removal and uniform distribution of flow. This was achieved by a dramatic reduction in the strength of the vertically impinging flow towards the belt. However, this "FD" type nozzle generated a dead zone near the triple point within the output chamber. To remove the dead zone, a gently sloped shape insulator was inserted between the tundish back wall and the belt.

High grade energy, which is primarily derived from hydrocarbon fuels, is in short supply; therefore alternative energy sources such as renewable and recycled energy sources are gaining significant attention. Pyro-metallurgical processes are large consumers of energy. They in return generate large quantities of waste heat which goes un-recovered. The overall theme of this research is to capture, concentrate and convert some of this waste heat to a valuable form. The main objective is to characterize and develop heat pipe technology (some of which originated at McGill) to capture and concentrate low grade heat. Heat pipe employs boiling as the means to concentrate the energy contained in the waste heat and transfers it as higher quality energy. The distinct design features of this device (separate return line and flow modifiers in the evaporator) maximize its heat extraction capacity. During the testing the main limitations within the heat pipe were identified. Different test phases were designed throughout which the configuration of the system was modified to overcome these limitations and to increase the amount of extracted heat. / L'énergie d'haut grade de nos jours est produite principalement à base de combustion d'hydrocarbure et les réserves de cette énergie deviennent de plus en plus rare, mais certaines énergies alternatives connues gagnent des forces parmi les marchés incluant les sources d'énergie renouvelables et recyclées. Les usines pyrométallurgiques sont des consommateurs significatifs d'énergie d'haut grade. Ces procédés industriels relâches un montant important de chaleurs (perte) à l'environnement sans aucune récupération. Le but du projet est de concentrer, capturer et convertir cette chaleur résiduelle de basse qualité en énergie valable. Par contre, l'objectif principal du projet comme tel est de développer et de perfectionner un caloduc capable d'extraire cette chaleur parvenant des gaz effluents. Le point d'ébullition d'une substance (vapeur) est utilisé comme moyen de concentrer l'énergie contenu dans les effluents avec la technologie des caloducs. Pour maximiser les gains énergétiques, la conception de ce caloduc en particulier utilise des canaux de retour indépendant ainsi qu'un modificateur de débit dans l'évaporateur, lui permettant d'extraire un niveau supérieur de chaleur. Pendant les essais lors du projet, les éléments limitants des systèmes de caloducs ont été identifiés. Les configurations du système ont été ajustées et modifiés dans la phase expérimentale d'essai pour surmonter ces limitations et maximiser l'extraction de chaleur.

Under certain conditions of moisture and oxygen, sulphides can spontaneously heat, known as self-heating or pyrophoric behaviour. In this thesis the hypothesis that galvanic interaction between some sulphides can promote self-heating is tested. Galvanic interaction is controlled by rest-potential difference between the minerals and the surface area of contact (particle size). In order of decreasing rest-potential, four sulphides were tested: pyrite, chalcopyrite, sphalerite and galena. Two series of samples consisting of mixtures of two sulphides each representing a mass fraction of 50 % were run in standard self-heating tests. The first series comprised five mixtures were prepared, two with low rest-potential difference (pyrite-chalcopyrite and chalcopyrite-sphalerite) and three with high rest-potential difference (pyrite-galena, chalcopyrite-galena and pyrite-sphalerite). The second series was performed on a pyrite-sphalerite mixture at four particle sizes (80 % passing 850 μm, 300 μm, 75 μm and 38 μm). The first series showed that the individual sulphides and the mixtures of low rest-potential difference did not self-heat but the mixtures of high rest-potential difference did self-heat. The second series showed that self-heating increased inversely with particle size (increasing specific surface area) and that it was the fineness of the pyrite (the high rest-potential sulphide) that governed the self-heating effect, indicating the rate-limiting reaction is reduction at the more noble pyrite (cathodic mineral). The increase in self-heating with high rest-potential difference and increasing particle fineness supports the hypothesis that galvanic interaction contributes significantly to sulphide self-heating. A possible mechanism based on the H2S hypothesis is proposed. The understanding gained will be of interest to those involved in storage, shipping and disposal of sulphide mineral mixtures. / Sous certaines conditions d'air et d'humidité, les sulfures peuvent s'auto-échauffer, et dans un cas extrême, subir une auto-combustion. Le présent mémoire étudie l'hypothèse qu'un effet galvanique entre certains sulfures pourrait promouvoir l'auto-échauffement. L'effet galvanique est contrôlé par la différence de potentiel de corrosion (ΔE) et l'aire de surface en contact (granulométrie). En ordre décroissant de potentiel de corrosion, les quatre sulfures étudiés sont: la pyrite, la chalcopyrite, le sulfure de zinc et la galène. Deux séries d'échantillons ont été soumises à des tests standards d'auto-échauffement. Les séries se composaient de mélanges de deux sulfures, chaque sulfure représentant une fraction massique de 50 %. Dans la première série, cinq mélanges ont été préparés, deux avec une différence de potentiel de corrosion de 0,1 volt (pyrite-chalcopyrite et chalcopyrite-sulfure de zinc) et trois d'un potentiel de corrosion > 0,2 volt (pyrite-galène, chalcopyrite-galène et pyrite-sulfure de zinc). Dans la deuxième série, un mélange de pyrite-sulfure de zinc à quatre granulométries différentes (80 % passant 850 μm, 300 μm, 75 μm et 38 μm) a été étudié. Dans la première série, les sulfures individuels et les mélanges possédant une différence de potentiel de corrosion de 0,1 volt ne se sont pas auto-échauffés, mais les mélanges ayant une différence de potentiel de corrosion de > 0,2 volt se sont auto-échauffés. La deuxième série étudiée montre que l'auto-échauffement augmente de façon inversement proportionnelle à la grosseur des particules (aire de contact plus grande) et que la finesse de la pyrite (sulfure au potentiel de corrosion le plus élevé) gouverne l'effet galvanique, indiquant que la réaction cinétiquement limitante est la réduction sur le sulfure le plus noble, donc la pyrite (la cathode). L'augmentation d'auto-échauffement dû à la différence de potentiel de corrosion > 0,2 volt et une granulométrie fine mettent en évidence l'hypothèse qu'un effet galvanique contribue d'une façon significative à l'auto-échauffement des sulfures. Un mécanisme basé sur l'hypothèse de l'acide sulfhydrique (H2S) est suggéré. Les résultats seront d'intérêt pour les personnes responsables de l'entreposage, du transport et de l'élimination de mélanges de sulfure.

A new thermal analysis technique was developed and tested. It makes use of the improvements in heat transfer characteristics associated with recent advances in heat pipe technology. Heat is extracted from a liquid sample of a melt taken in-situ from within a vessel or furnace. The rate of heat extraction is such as to cause the sample to solidify. The technique was tested both in the laboratory and on an industrial scale (Grenville Castings, Perth, Ontario). Aluminum alloys including 356, 319, Al-xSi, Al-Si-Cu-xMg, and 6063 were subjected to various melt treatments and were used to carry out the tests. Classical thermal analysis was also carried out simultaneously under the same melt conditions using a preheated graphite cup. / The comparison showed that the new technique has great potential over classical thermal analysis. The major advantages of the new method are that it conducts the analysis inside the melt (since it is no longer necessary for a physical sample to be removed from the melt itself), it consumes less time and the cooling rate can be precisely controlled during the solidification process. Moreover, it produces curves of greater detail and of better resolution than conventional techniques. In fact, the detail is of such resolution that, in some cases, the cooling curves may be used to infer the chemical composition of certain components of the melt, a fact which equates to a form of rapid chemical analysis. The peaks in the signal which refer to intermetallic formation are of better resolution and more identifiable when the new technique is used. The size of the peaks obtained using the new probe is about three times greater than that obtained by the classical method. With this new technique it becomes possible to correlate the area below the intermetallic peak to the concentration of iron or copper in the melt. This is a feature which makes the new thermal analysis probe act as a rapid chemical analyzer for selected constituents.

Osprey process is a new kind of metal forming technology. In this process, a stream of liquid metal is atomized into a spray of molten droplets by a high velocity inert gas jet. The atomized droplets are accelerated towards a substrate of suitable shape and size. At the same time, they are rapidly cooled by the surrounding relatively cold gas and thereby partly solidified. After a certain flight distance, the droplets impinge on the cold substrate, and solidification continuous on the substrate. Near-net shaped products can be manufactured by this process. / In this study, a simple mathematical model is established to describe the atomizing gas velocity profile and the velocity, thermal and solidification profiles of rapidly cooled metal droplets of different sizes during the in flight droplet-gas interaction. Given the relevant spray parameters, the model allows to predict quickly the transient droplet velocity, temperature, and solid fraction contents of individual droplets at various spray distances from the substrate. This model can be used to ascertain the suitability of the nozzle-substrate distance in Osprey process. The developed mathematical model has been used to predict thermal history and solidification behavior of atomized droplets of gamma-TiAl alloy. The model predicts undercooling, nucleation temperature, nucleation position and the extent of solidification of the in flight droplets of sizes ranging from 20 mum to 500 mum.

Physical and mathematical modelling studies were performed, in order to analyze various transport phenomena occurring during steel making tundish operations. Their effects on liquid metal quality were reported. A full-scale water model of a twelve tonne, delta shaped, four strand, billet caster tundish was used for physical modelling. The commercial code ANSYS FLUENT 12 was used for carrying out mathematical modelling. The tundish used in the present study is a full scale replica of that operated at the RTIT/QIT plant in Sorel Tracy, Canada and is located at MMPC's water modelling laboratory at McGill University. It is a long lasting fact that the flow pattern within a tundish greatly affects the output metal quality. As such the insertion of flow modifiers in a tundish is a common practice. In the present study, eighteen different arrangements of flow modifier systems (combinations of impact pad and dams) were considered, and mathematical modelling was performed to predict the inclusion removal efficiency for each tundish configuration. A new dimensionless number (Gu) has been proposed, which is a good measure of steel cleanliness. During melt transfer from the ladle to the tundish, inert gas is injected into the ladle shroud, just below the slide gate, so as to prevent aspiration of ambient air. The effect of inert gas shrouding on the fluid flow patterns and slag movements have been numerically predicted by using a 3D mathematical model, and then validated with water model experiments. The effect of the alignment of the ladle shroud during melt transfer was also studied, using a 3D mathematical model, supported by subsequent water model experiments. It was demonstrated that a slight bias from the vertical can be very detrimental to steel quality. Remedial measures have been suggested. During typical steelmaking tundish operations, conditions are generally non-isothermal. Variable heat losses take place from the free surface and from the walls of the tundish. Similarly, during a ladle change, the steel poured in from the new ladle will tend to be at a higher temperature than the liquid steel remaining in the tundish. Flow patterns change under non-isothermal conditions and hence affect output steel quality. A thorough study has been performed to visualize the effect of thermal gradients on fluid flow patterns, and temperature distributions generated within the delta shaped tundish. / Dans la présente étude, la modélisation physique et numérique fut utilisée pour analyser l'effet sur la qualité du métal liquide de différents phénomènes de transports ayant lieu dans un panier répartiteur durant les opérations de coulée d'acier. Un modèle physique pleine échelle d'un panier répartiteur de forme triangulaire d'une capacité de douze tonnes comprenant quatre jets pour la coulée de billettes ainsi qu'un modèle mathématique utilisant le logiciel ANSYS FLUENT 12 fut utilisé. Le panier utilisant l'eau comme fluide plutôt que l'acier est une réplique de celui utilisé à l'aciérie RTIT/QIT de Sorel Tracy, Canada. Il est bien connu que les patrons d'écoulement qui se développent dans le panier affectent grandement la qualité du métal à la sortie et, en conséquence, l'insertion de modificateurs d'écoulements est pratique courante. Dans la présente étude, dix huit arrangements différents de modificateurs d'écoulements (panneaux d'impacts et digues) furent considérés et furent numériquement modélisés pour prédire l'efficacité à limiter l'entrainement d'inclusion lors de la coulée. Durant le transfert de l'acier de la poche de coulée vers le panier répartiteur, du gaz inerte est injecté dans le jet immédiatement en dessous de la valve coulissante pour prévenir l'aspiration d'air ambiant. L'effet de l'injection du gaz sur le patron d'écoulement et sur les mouvements du laitier de surface a été estimé grâce à un modèle mathématique en volume (3D) et les résultats furent validés expérimentalement en utilisant le modèle physique à l'eau contenant un laitier de microbilles de verre creuses flottantes. L'effet de l'alignement du jet provenant de la poche de coulée qui alimente le panier répartiteur fut aussi étudié par modélisation mathématique en trois dimensions et subséquemment, par des tests physiques avec l'eau. Il fut prouvé que les conséquences d'un léger désalignement vertical du jet est catastrophique et des solutions correctrices sont proposées.Durant les opérations réelles de coulées de l'acier avec panier répartiteur, les conditions ne sont pas isothermes. Il y a des pertes calorifiques provenant de la surface et des côtés du panier. Durant les opérations de changement de poche de coulée, l'acier provenant de la nouvelle poche peut aussi être plus chaud que le restant du panier provenant de la poche précédente. Les patrons d'écoulement change donc sous ces effets non isothermes et affectent aussi la qualité de l'acier sortant. Une étude poussée fut menée pour illustrer l'effet des gradients thermiques sur les patrons d'écoulements et sur la distribution de température dans le panier triangulaire.