A numerical study was carried out to predict three dimensional turbulent flow heat and mass transfer in RH-ladles. To model the turbulence in the system, the popular Launder and Sharma (7) version of the low Reynolds number k-$ varepsilon$ model was employed. The governing equations for fluid flow and heat transfer were solved using a control volume-based finite difference method. The buoyancy effect in the axial momentum equation was incorporated through the Boussinesq approximation. From the fluid flow and turbulence characteristics predicted by the numerical model, the mass transfer occurring from the dissolution of steel rods immersed in the liquid steel was computed using the Sherwood number correlation recently proposed by Mazumdar et al. (25). The parameters studied include: the flow rate of steel, the heat flux from the solid walls and various radial positions of the vertically immersed steel rod. The fluid flow results are presented as vector plots. The heat transfer aspects are shown through temperature contours at various vertical planes and the mass transfer results are graphically presented in plots showing variance of mass transfer coefficients with key parameters. / The flow field results were found to be in qualitative agreement with previously reported 3D numerical studies for similar systems. Due to the lack of any experimental or numerical results related to heat and/or mass transfer in RH-ladles, the heat and mass transfer results obtained in the present study could not be compared and verified.

A series of experiments determined the influence of oxygen partial pressure and SiO$ sb2$ content in soda slag on the distribution of metals between soda slag and dore metal. Oxygen partial pressure ranged from 10$ sp{-8}$ to one atmosphere and the SiO$ sb2$ content from 0 to 10 wt%. / Minor elements, including the precious metals, partition between the dore metal and the soda slag. The main purpose of this research was to determine the process parameters which influence this distribution, and the nature and forms of the precious metals in the metal and in the slag phases. / From the results of the experiments it was found that: (a) Increasing oxygen partial pressure from 10$ sp{-8}$ to 1 atm does not influence gold, palladium and platinum removal from dore metal, but leads to the practically complete removal of Te (down to 4.8 ppm), Rh (down to 2.7 ppm), Pb (down to 6.0 ppm) and quite good removal of Se (down to 0.0112 wt%) and Cu (down to 0.42 wt%) from the dore metal. (b) Silica content of the soda slag does not influence gold, palladium and platinum chemical losses with soda slag. However it was found that increasing silica in slag leads to an increase of mechanical loss of precious metals because of the increase in slag viscosity. (c) Increase in silica content of slag increased lead and copper removal from dore metal, and decreased the removal of rhodium, selenium and tellurium, especially at low oxygen potential. (d) It was concluded that most of precious metals leave dore metal in the form of mechanical losses, but silver and rhodium were found in slags as intermetallic phases with the other components. Rhodium is also lost in the form of the volatile oxide RhO$ sb2$. (Abstract shortened by UMI.)

The effects of austenite grain size, level of undercooling, and strain and strain rate in compression on the austenite-to-ferrite transformation in a 0.1% C, 1.4% Mn steel were investigated. The influence of this transformation on the hot ductility of the steel in question was then examined. Samples were strained at rates from $ rm 3 times 10 sp{-4} s sp{-1}$ to $ rm 3 times 10 sp{-2} s sp{-1}$ after cooling to holding temperatures between the Ae$ sb3$ and the (undeformed) Ar$ sb3$ temperature. / Deformation was found to raise the effective Ar$ sb3$ temperature almost to the Ae$ sb3$, and to significantly accelerate the transformation in both fine ($ sim$25 $ mu$m) and coarse ($ sim$200 $ mu$m) grained samples. At relatively small levels of undercooling ($ sim$40$ sp circ$C, it was found that deformation induced ferrite could not form outside the regions of the material that were significantly affected by the applied strain. At low strain rates and in coarse grained material, this meant that ferrite formed in narrow bands near the grain boundaries. No such restriction was found in fine grained material. / In tension, it was found that, if thin bands of grain boundary ferrite were present in the microstructure, failure always occurred in this second phase. The effect of intergranular ferrite on the ductility was found to vary with the amount present; the lowest %RA value was obtained with around 10% ferrite present in the microstructure. The detrimental effect of ferrite lessened slightly as the strain rate was raised. / It is concluded that, from an industrial perspective, unless unbending can be performed at strain rates in excess of $ rm 3 times 10 sp{-2} s sp{-1},$ it would be prudent to carry out this operation either above the Ae$ sb3$ or below the (undeformed) Ar$ sb3$ temperature for the steel in question.

A new fast response conductivity meter was developed and tested. The "five time constant" of the meter is 0.08 s which meets the requirement for measurements under the dynamic conditions relevant to this work. / In a laboratory column, a bubble interface was created by introducing a step change of gas flow, and the rising velocity of this interface, $u sb{in},$ was measured using a conductivity method with the new conductivity meter. A measurement of the three-dimensional bubble swarm velocity in the column was obtained by interpolation from the $u sb{in}$ measured as a function of $J sb{g2} vert J sb{g} sb1 ,$ where $J sb{g} sb1$ and $J sb{g} sb2$ are the superficial gas velocities before and after a step change of gas flowrate, respectively. This velocity was referred to as the hindered velocity, $u sb{h}.$ The buoyancy velocity, $u sb0 ,$ was readily determined by switching off the gas, i.e. $u sb0 = u sb{in}$ at $J sb{g} sb2 = 0.$ / The average gas velocity, $u sb{g},$ was corrected to the local average gas velocity, $u sb{g,loc},$ to obtain the average gas velocity under the local pressure conditions at a given vertical position in the column. The experimental results showed that $u sb{h}$ was significantly less than $u sb{g,loc}$ (and $u sb{g}).$ This is because the $u sb{h}$ is the three-dimensional bubble swarm velocity and $u sb{g,loc}$ is the one-dimensional bubble swarm velocity. Unlike $u sb{g,loc},$ the $u sb{h}$ was constant along the column, which was supported by theoretical momentum analysis. The $u sb{h}$ is proposed as the key characteristic swarm velocity of the system. / For the air-water only system in the two-dimensional domain, using parabolic models for gas holdup and liquid circulation velocity profiles over the cross section of the column, the $u sb{h}$ could be fitted to the experimental data. For the air-water-frother system, the $u sb{h}$ could not be fitted to the experimental data which is attributed to the air bubbles adopting a circulatory flow pattern. / In the air-water only system under batch operation, Nicklin's derivation (1962), i.e. $u sb{g} = u sb0 + J sb{g},$ was supported only under restrictive conditions, namely $u sb{g}$ and $J sb{g}$ must be measured at atmospheric pressure. Considering the local values, the experiments showed that $u sb{g,loc}$ was not equal to $u sb0 + J sb{g,loc}.$ In the presence of frothers under batch or countercurrent operation, the experiments showed that Nicklin's derivation was not applicable even if atmospheric values of $u sb{g}$ and $J sb{g}$ were used.

A common wastewater treatment process practiced by zinc production facilities is the single-stage mixing of acidic wastewaters with slaked lime, inducing the reactive precipitation of fine (∼1 mum) gypsum (CaSO4.2H 2O) and other solids with a solids density less than 10%. These solids report to a tailings pond for containment. / Tailings pond life would be increased if the solids density of the precipitated solids was improved. Previous work at McGill University suggested that a staged neutralization process with solids recycle and seeded with gypsum would produce large-sized gypsum crystals with a high solids density. A continuous lab-scale process run with synthetic zinc plant effluent produced large (∼100 mum) gypsum crystals with a solids density of 50 +/- 3%. / Meissner's method of calculating mean activity coefficients allowed for the calculation of gypsum solubility in mixed, strong sulphate electrolyte solutions.

Arsenic is a major contaminant in the non-ferrous extractive metallurgical industry and its disposal poses a serious environmental threat. Precipitation of arsenic from aqueous acidic solutions in the form of crystalline scorodite (FeAsO4.2H2O), which is a naturally-occurring mineral, is proposed as an arsenic fixation method. The production of scorodite is usually done in autoclave at high temperature and pressure. However, in a breakthrough at McGill University, a step-wise neutralization procedure has been developed where scorodite can be formed at ambient pressure (95°C). This work investigates the relation between precipitation conditions in sulphate media with the crystallinity, yield, and above all, solubility of scorodite produced. This work has investigated the effect of (1) type of seed, (2) type of base, (3) cationic and anionic additives, (4) Fe/As ratio, and (5) presence of Fe(II) in the precipitation media. It was found that crystalline scorodite can be produced via the use of various types of seed such as hematite and gypsum in addition to scorodite. The stability of the product was not found to be affected measurably upon change of base from MgO to CaO and in the presence of various cations (Cu, Zn, Ni, Co, and Mn) and anions (sulphate, nitrate and phosphate). Most solubilities measured were between 1 and 3 mg/L at pH 5. Solubilities as low as 0.5 mg/L were found when Fe(III):As(V) was 3:1 but the yield was low. Following these results, Fe(II) was added to the system and solubilities as low as 0.2 ppm were observed with yields of over 80% in a single precipitation step.

A thickener is a continuous gravity separation device that reduces or removes suspended solid particles from liquor. Clarified liquor is removed from the top and thickened solids are discharged from the bottom. Thickeners are an essential part of plant water management. / A conductivity-based sensor has been developed for use in a thickener and has been successfully tested in industrial applications. Data from at Falconbridge's Kidd Creek operations, Inco's Thompson and, in particular, Inco's Sudbury operations are discussed. / The probe designed for this work is a multi-cell arrangement that exploits the difference in conductivity between the liquor and the slurry. Conductivity measurements are taken as a function of depth to provide a profile of the solids content of the thickener. Conductivity measurements can be converted to solids concentration (percent solids) using a model developed by Maxwell. The resulting solids concentration versus depth profile was used to interrogate the behaviour of the solids and to develop thickener control signals. (Abstract shortened by UMI.)

The quality of the strip produced in a near-net-shape casting process is strongly linked to the metal feeding system. An alternative metal delivery system was proposed and a comprehensive mathematical model, including heat transfer, fluid flow and stress studies was used as a tool to perform this task. This model included three-dimensional flows with fully coupled momentum and heat transfer. Turbulence effects as deduced from the standard kappa- 3 equations, macroscopic solidification, as well as flow through porous media were also modeled. A further simple model for stress was employed to calculate the displacements and the stress distribution throughout the forming strip. / Besides the mathematical model, an experimental apparatus was developed, reproducing as closely as possible the type of metal-substrate contact occurring in a single-belt caster. Casting of thin strips at high speeds was achieved and a parametric study was carried out. / The standard extended nozzle configuration did not provide a smooth flow in the reservoir. Recirculation zones besides a nonuniform delivery of the metal to the belt were observed. Divergences in the solid fractions between the center and the edges of the strips were found to increase with increasing turbulence at the exit gap. The values of the morphology constant C and the overall heat transfer coefficient h were considered the most important inputs of the model. Furthermore, modelling turbulence through a kappa- 3 model proved to be critical for this particular configuration. / The adoption of a flow modifier within the extended nozzle eliminated the recirculation zone in the reservoir and yielded a smoother flow to the belt. Such improvements in the flow distribution were decisive to ensure more uniform solidification along the width of the strip. / In early solidification, the strip was found to displace inwardly, due to the cooling and contraction of the solid shell. Tensile stresses were calculated at the cooling surface, whereas compressive stresses were found in the middle of the strip. / Interfacial heat fluxes and heat transfer coefficients measured on the strip casting simulator were highly dependent on the contact metal-substrate-coating. Values of q and h varied threefold for the various coatings and levels of roughness investigated. Correlations were derived to predict peak heat fluxes and the evolution of q after the peak, for the various coatings. A relationship was observed between the peak heat transfer coefficient and the total solidification time. (Abstract shortened by UMI.)

Strip casting is a new process, which can produce cast strips with thicknesses of only 2--5mm and eliminate hot rolling. Cold rolled strips produced by strip casting have low mechanical anisotropy, which is related to the texture of the strip; however, this link has not yet been investigated. Texture development from strip casting through to cold rolling and subsequent annealing was therefore studied. / Cast strips of two types were produced by solidification into the dual (delta ferrite + gamma austenite) phase and single (delta ferrite) phase domains. The {001}⟨uvw⟩ delta ferrite texture is converted into the {001}⟨uvw⟩ gamma final texture in the first steel, while most of it transforms into the fcc {011}⟨uvw⟩ fiber in the second. The observed variant selection can be explained using the Bain strain model. / The cast strips and a hot rolled strip were subsequently cold rolled and annealed. The gamma-to-alpha' deformation-induced transformation took place during cold rolling and low symmetry rolling textures were produced in the (delta ferrite + austenite) steel after 50% reduction by cold rolling. By contrast, the 80% cold rolled textures displayed orthorhombic symmetry. The recrystallization texture components changed from {001}⟨100⟩ and {110}⟨111⟩ to {110}⟨001⟩, {110}⟨112⟩ and {112}⟨111⟩ when the cold reduction was increased. The cast strips displayed weaker recrystallization textures than the conventionally processed material. Moreover, the delta ferrite steel contained stronger {110}ND fiber components than the two-phase material. Texture development in the rolled and annealed products is discussed in terms of the Kurdjumov-Sachs transformation relationship.

Experiments were carved out in a stirred tank reactor with both two phase and three phase systems. A systematic examination of the effect of solid particles on parameters such as volumetric liquid-side mass transfer coefficient kLa, specific interfacial area a, gas holdup epsilong, bubble size dB, and average bubble residence time tau, was conducted. Both dimensional and dimensionless correlations were generated and conclusions were drawn. / The work was carried out in the context of the industrial experience that stirred tank reactors are widely used in both metallurgical and chemical engineering processes that often involve gas absorption in two and three phase systems. The effect of solids loading on the oxygen absorption rate in stirred tank reactors is complex due to the interactions between bubbles and solid particles, and also between other parameters. The measurement of some parameters, such as bubble size, specific interfacial area, volumetric mass transfer coefficient, is difficult, and there are many discrepancies in the literature with respect to the form of the correlation for oxygen absorption rate and bubble size and also the explanation of the internal mechanisms. / The present work found that kLa decreased with increasing solids volume fraction. This decrease was as much as 50% when solids loading was increased from 0 (water) to 14 vol% at a superficial gas velocity of 0.64 cm/s and a impeller speed of 400 rpm. The primary reason for this was due to the enhancing effect of solids on bubble coalescence and the increase of slurry density. Kolmogroff's turbulence theory was proved not applicable to the present STR and conditions. The cause for this was believed to be due to the fact that the power density in the STR was far from homogeneous, and the dynamic equilibrium between bubble breakup and the bubble coalescence was not established. / Bubble size, dB, was found to be increased by increasing solids volume fraction. The increase was due to the direct enhancing effect of solids on bubble coalescence and due to less bubble recirculation, i.e., change of the bulk flow pattern, with increased solids loading. There was no direct relationship found between a change of the apparent viscosity of slurry and a change of bubble size. Therefore, the manner by which many people incorporated apparent viscosity into the correlation of dB is inappropriate. / To properly describe the extent of bubble coalescence in three phase systems, a new parameter, the coalescence factor, fc, was defined in the present work. It was found that the coalescence factor decreased with increasing solids volume fraction and power input, but decreased with increasing superficial gas velocity.