Thesis (DEng (Process Engineering))--University of Stellenbosch, 2005. / This dissertation covers research carried out over the past 25 years in the area of flotation. Flotation is one of the most widely used processes in the beneficiation of minerals. The process is characterized by a number of important sub-processes each with their own set of critical variables. These include: • The pulp phase which is influenced by variables such as pH, Eh, the nature of the chemical reagents used, the chemical state of the surface of the ore particles, etc. • The reactor in which the process occurs, viz. the flotation cell, which is influenced by factors such as aspect ratio, degree of agitation, mechanical design criteria, energy input, aeration processes, etc. • The froth phase which is arguably the heart of the process and probably the least well understood but which is influenced by factors such as size and shape of the solid particles in the froth, the nature of the surfactant used, the aeration rate, the water recovery rate, the froth depth, etc.

Thesis (MScIng)--University of Stellenbosch, 2003. / ENGLISH ABSTRACT: Spherical wax particles with diameters of 5 μm and less are known to enhance the properties of powder and solvent-based coatings. Conventional methods, including spray freezing and jet milling, yield only 15 – 20 % of particles within this desired size range. Recently, supercritical fluids have been used in the micronisation of a variety of polymers, pharmaceuticals and other inorganic solids. Of these processes, the Rapid Expansion of Supercritical Solutions (RESS) involves the expansion of a solid dissolved in a supercritical fluid over a nozzle. In the process the solubility of the solid is drastically decreased to yield high supersaturated conditions, resulting in a multitude of small and stable wax nuclei. The technique is known to produce small, solvent free particles with a narrow size distribution. In this project, the RESS technique was applied to two commercial Fischer-Tropsch waxes: Paraflint C80 and Paraflint C105. Propane was selected as supercritical solvent due to its increased solvent power for paraffin wax compared to other conventional supercritical solvents such as carbon dioxide and ethane. From theoretical studies, nozzle geometry, as well as nozzle inlet temperature, pressure and concentration was identified as the parameters most likely to influence the size and shape of the wax particles. Porous plate nozzles were chosen as expansion device due to lower pressure drop and higher flow rate of the expanding solution compared to conventional capillary nozzles. Pre-expansion conditions of 125 to 160 ºC and 125 – 180 bar were investigated. Wax concentrations up to 5 weight % were expanded through porous plate nozzles with pore sizes of 15, 25 and 60 μm. It was found that small particles, less than 5 μm in size could readily be produced for the whole range of parameters investigated. Average particle sizes of less than 2 μm were found in virtually all cases. The C80 particles tended to be less smooth and of more angular shape than C105 particles. It is argued that the higher fraction of low weight compounds softens the C80 wax particle surface, making them more susceptible to deformation and coagulation. Experimental studies confirmed that a combination of lower temperature and higher pressure (175 bar and 130 ºC) increases the chances for the formation of spherical particles. This combination corresponded to conditions furthest away from the dew point of the solution. The 25 μm porous plate nozzle gave better results than the other nozzles, indicating that the characteristics of the nozzle could play a more important role than the pore size. SEM images indicated that this nozzle has a larger porosity, which could have resulted in lower pressure drops compared to the other nozzles. Higher wax concentration did seem to give smaller individual particles, but particle shape visibly deteriorated as the concentration increased. From experiments with C105 wax, at these optimal conditions, it seems as if a maximum concentration between 4 and 5 % is achievable, above which the particle shape became flak-like with rough edges. Due to limitations of the batch experimental set-up, especially regarding the reaching of steady state, it is strongly recommended that the obtained results be verified on pilot plant scale, where longer spray times can be achieved. The RESS process for wax micronisation on industrial scale needs to be critically examined and higher pressure and solvent recycling costs would need to be weighed up against the lower wax recycling required to obtain the desired particles. / AFRIKAANSE OPSOMMING: Sferiese was partikels met ’n deursnee van 5 μm en minder is bekend daarvoor dat dit die eienskappe van ’n verskeidenheid van poeier en oplosmiddel-gebaseerde deklae verbeter. Alledaagse metodes, waaronder sproeivries- en maalprosesse soos straalmaling, lewer slegs sowat 15 – 20 % van die partikels binne die gewenste partikelgrootte. In die afgelope tyd is superkritiese vloeistof mikroniseringsprosesse gebruik om ’n verskeidenheid van polimere, farmaseutiese en ander anorganiese stowwe te mikroniseer. Die sogenaamde “Rapid Expansion of Supercritical Solutions (RESS)”– proses behels die ontspanning van ’n oplossing van die vaste stof en ’n superkritiese oplosmiddel oor ’n spuitstuk. In die proses verlaag die oplosbaarheid van die vaste stof drasties, sodat hoë superversadigde vlakke in ’n baie kort tydperk bereik word om ’n menigte klein en stabiele was kerne te vorm. Die tegniek is bekend daarvoor om fyn, oplosmiddelvrye partikels met ’n nou partikelgrootteverspreiding te lewer. In hierdie projek is die RESS-proses ondersoek om twee Fischer-Tropsch wasse, C80 en C105, te mikroniseer. Propaan is as oplosmiddel gekies weens die verhoogde oplosbaarheid van paraffienwasse daarin vergeleke met ander konvensionele superkritiese oplosmiddels soos koolsuurgas en etaan. Poreuse skywe is as spuitstukke verkies weens die feit dat hoër vloeitempos en laer drukvalle oor hierdie spuitstukke verkry kan word. Na ’n teoretiese ondersoek is spuitstukgeometrie, asook die temperatuur, druk en konsentrasie by die spuitstukinlaat geïdentifiseer as die prosesparameters wat die partikelgrootte en –verspreiding die waarskynlikste sal beïnvloed. Daar is besluit om die spuitstukinlaatkondisies van 125 tot 160 ºC en 125 tot 180 bar te ondersoek. Waskonsentrasie tot en met 5 massa % is deur poreuse skywe met gemiddelde poriegroottes van 15, 25 en 60 μm ontspan. Klein waspartikels, met gemiddelde partikelgroottes van minder as 2 μm is in feitlik al die gevalle verkry. Die C80 waspartikels het geneig om minder glad en meer hoekig as die C105 partikels te wees. Hier word vermoed dat die hoër hoeveelheid lae massa komponente die C80 waspartikeloppervlak versag en dit meer vatbaar maak vir vervorming en koagulasie. Eksperimentele studies het getoon dat ’n kombinasie van laer temperature en hoër drukke (175 bar en 130 ºC) by die spuitstukinlaat die kanse vir sferiese partikels vergroot. Hierdie temperatuur/druk kombinasie val saam met toestande verder weg van die doupunt van die oplossing. Sover dit die spuitstukke aanbetref, het die 25 μm poreuse skywe beter resultate as die ander spuitstukke gelewer. Hier word vermoed dat die fisiese eienskappe van die spuitstuk ’n meer prominente rol as die poriegrootte speel. SEM foto’s het getoon dat hierdie spuitstuk ’n groter porositeit, wat ’n laer druikval tot gevolg kon gehad het, in vergelyking met die ander spuitstukke. Dit blyk asof hoër waskonsentrasies lei tot kleiner individuele partikels, maar die vorm en oppervlak daarvan het sigbaar verswak by hoër konsentrasie. Vanuit eksperimente met C105 was, by die bovermelde optimale kondisies, wil dit voorkom asof ’n maksimum konsentrasie tussen 4 en 5 % haalbaar is, voordat die partikels ’n vlokkierige vorm met ruwe oppervlak aangeneem het. Weens die beperkinge van die eksperimentele opstelling, word dit ten sterkste aanbeveel dat die resultate op loodsaanlegskaal, waar langer sproeitye moontlik is, bevestig word. Die RESS-proses behoort krities ondersoek te word en die hoër druk en oplosmiddel hersirkuleringskoste moet opgeweeg word teenoor die voordele van die laer was hersirkulering benodig om die gewenste partikels te lewer.

Thesis (PhD (Process Engineering))--Stellenbosch University, 2008. / The feasibility of predicting paint properties directly from the raw material formulation as well as the rheological data is investigated in this study. Although extensive work has been carried out on the prediction of paint properties in terms of the raw material data, very little research has been carried out on the prediction of paint properties in terms of the rheological data. Little is known about the relationship between fundamental rheological properties and real-world performance. The paint under investigation consists of fourteen raw materials. These raw materials interact in a very complex manner to produce certain desired paint properties. Evaluation of these interactions in terms of constitutive equations is almost impossible and the relationships between paint properties, raw materials and rheology can only be modelled in a statistical way. Linear relationships are investigated with linear parameter estimation techniques such as multiple linear regression. However, it has been found that many of these relationships are non-linear and that linear modelling techniques are no longer applicable for certain situations, e.g. at very high concentrations of specific raw materials. Non-linear techniques such as neural networks are used in these situations. The relationship between the raw materials, paint properties and rheology are evaluated using the following three models: · MODEL 1: The relationship between rheology and raw materials · MODEL 2: The relationship between paint properties and raw materials · MODEL 3: The relationship between paint properties and rheology MODEL 1 makes use of techniques such as principal component analysis and preliminary modelling to respectively reduce redundancy and to capture as much data as possible. MODELS 2 and 3 make use of linear screening techniques in order to identify relevant raw materials and paint properties. The validity of every model is checked to ensure that predictions and interpretations are unbiased and efficient. MODEL 1 revealed that emulsion, extender particles, pigment, water, organic pigment and solvent are the six most important raw materials affecting the rheology of the specific paint. The rheology curves that are predicted most accurately by means of multiple linear regression are the “Amplitude Sweep” (AS), “3-Interval-Thixotropy-Test” (3-ITT) and the “Flow Curve” (FC). Non-linear rheological behaviour is encountered at high pigment volume concentrations (PVC) and volume solids (VS), due to the strong dependency of the rheology of the paint on these properties. It has been shown that neural networks perform better than multiple linear regression in predicting the rheological behaviour of these paint samples for which the raw materials vary by more than 20% from the standard formulation. On average, neural networks improve predictability of the rheological parameters of these samples by 54%. The largest improvement in predictability is made on the rheological variable “Extra Low Frequency” value (CXLF), where multiple linear regression resulted in relative errors of 59%, while neural networks resulted in errors of only 5%. Other predictions of rheology curves where neural networks have shown a major improvement on predictability are the “Time Sweep” (TS) – 68% increase in accuracy and “Low Shear” curve (LS) – 63% increase in accuracy. The smallest increase that the neural network had on the predictability of a rheology curve, was a 33% increase in accuracy of the “Amplitude Sweep” (AS) predictions. Multiple linear regression models of MODEL 2 predict the critical paint properties of Opacity, Gloss, Krebs Viscosity and Dry Film Thickness with relative errors smaller than 10%. It has been shown that 90% of all new predictions fall within the allowable error margin set by the paint manufacturer. Paint properties that can be predicted with an expected error of between 10% and 20% are Dry and Wet Burnish, Open Time and Water Permeability. Paint properties that are predicted the most inaccurately by MLR, that results in errors larger than 20% are Dirt Pick-Up and Sagging. Non-linear techniques such as neural networks are used to predict the paint properties of these paint samples for which the raw materials vary by more than 20% from the standard formulation. The neural networks show a major improvement on the predictability of the paint properties for those paint samples that vary more than 20% from the standard formulation. On average, neural networks improve predictability of the paint properties by 47%. The largest improvement in predictability is made on the Wet Burnish20 prediction, where multiple linear regression resulted in relative errors of 66%, while neural networks resulted in errors of only 0.6%. Other paint property predictions where neural networks have shown a major improvement on predictability of 80% or more in accuracy are Gloss – 80% increase in accuracy and Dry Film Thickness – 81% increase in accuracy. The smallest increase that the neural network had on the predictability of a paint property, was a 33% increase in accuracy of the Sag predictions. MODEL 2 makes it possible for the paint manufacturer to test tolerances around certain paint properties during manufacturing. Rheology is still a very under-utilised tool for explaining certain paint properties. MODEL 3 quantified the correlation between fundamental rheological properties and real world performance of a paint. It has been shown that rheological measurements can be used accurately to predict certain critical paint properties such as Opacity, Krebs Viscosity, Dry Film Thickness and Gloss within the allowable error margin given by the paint manufacturer. Multiple linear regression models predict the paint properties of Opacity, Krebs Viscosity and Dry Film Thickness with relative errors smaller than 10%, with rheology as input to the model. A neural network of MODEL 3 was developed to predict the paint properties of those paint samples that vary more than 20% from the standard formulation, by using rheology data as input to the model. The neural networks perform better than multiple linear regression in predicting the paint properties of these paint samples. On average, neural networks that use rheology data as input, predict the paint properties 49% more accurate than equivalent multiple linear regression models. The greatest improvement in model predictability is for Water Permeability - 73% increase in accuracy and Gloss - 70% increase in accuracy....

Thesis (PhD (Process Engineering))--University of Stellenbosch, 2006. / Recent advances in inorganic material preparation for membrane fabrication have extended the use of membranes to high temperature and chemically harsh environments. This has allowed inorganic membranes to be integrated into catalytic reactors, resulting in the concept known as Catalytic Membrane Reactors (CMRs). CMRs have overall important benefits of product quality, plant compactness, environmental impact reduction and energy savings. It has found use in a broad range of applications including biochemical, chemical, environmental and petrochemical systems. In these CMRs, the membranes perform a variety of functions, and consequently they are categorized according to the primary role of the membrane: extractor, distributor or contactor. In this dissertation the different uses of an extractor Catalytic Membrane Reactor (eCMR) are evaluated with the help of model reactions. In the eCMR the primary function of the membrane is to selectively extract one of the reaction products from the reaction zone, thereby combining the benefits of separation and reaction in one unit operation. This can lead to a number of advantages, of which the two most important ones include: (a) conversion beyond thermodynamic equilibrium in equilibrium restricted reactions and/or (b) the improvement of product selectivity in consecutive/parallel reaction networks. The dehydrogenation of isobutane, an equilibrium restricted reaction, was evaluated in a dense Palladium and a MFI-zeolite/alumina composite eCMR. These two eCMRs, consisting of a membrane packed with a Pt/In/Ge-MFI-zeolite catalyst, differed only on the basis of the membrane used. The palladium membrane showed superior extraction and selectivity capability for hydrogen in the reaction mixture compared to the MFI/alumina composite membrane. Regardless of these facts, the performances of the Pd and MFI eCMR, when evaluated at the same reaction conditions, were similar. The isobutane conversion to isobutene, employing high sweep rates (185 ml/min) could be increased up to ca. 37 % at 723 K, compared to 14 % in the conventional packed-bed reactor. The similar performance of the two different eCMRs was evaluated using a Catalytic Membrane Reactor model. Model results showed that in order for the extractortype CMR to completely draw benefit from the combination of membrane and catalyst in the same unit for conversion enhancement, a very active catalyst should be developed, able to sustain the high extraction ability of the membrane. This was the first time that these two eCMRs were evaluated at similar reaction conditions in order to study the influence of the nature of the membrane material on the working of the eCMR. The eCMR was also used to carry out meta-xylene isomerization. This part focused on the extraction of para-xylene from the meta-xylene isomerization reaction zone with a MFI eCMR (MFI-zeolite membrane and Pt-HZSM5 fixed-bed catalyst) in order to improve the reaction selectivity towards para-xylene. Para-xylene is an important industrial chemical used as a precursor for polyester resin, and in order to meet the paraxylene demand, ortho- and meta-xylenes are converted via the xylene isomerization reaction to xylene isomers. It has been shown that the pore-plugged MFI-zeolite membranes used in this study can selectively extract para-xylene from a mixture of xylenes. Using an extractor type catalytic membrane reactor instead of a conventional fixed-bed reactor for meta-xylene isomerization, can lead to higher para-xylene selectivities. The para-xylene selectivity can even be improved to 100% if the CMR is operated in the permeate-only mode, but this comes at a price of lower para-xylene yields. When operated in combined mode (i.e. mixing both permeate and retentate streams after the reactor), the CMR shows an improvement on both para-xylene productivity (ca. 10 % maximum at conditions studied) and selectivity when compared to the conventional reactor. This is the first time paraxylene selectivity could successfully be improved by employing an extractor Catalytic Membrane Reactor. This dissertation also led to the design and construction of a new generation membrane reactor testing bench, a first in the Department of Process Engineering, University of Stellenbosch. The bench allows for high temperature evaluation of membranes and Catalytic Membrane Reactors. The design is simple and easily adaptable for use to evaluate various different reactions.

Thesis (PhD (Process Engineeering))--University of Stellenbosch, 1995. / The hydrometallurgical leaching of sulphide concentrates was introduced in the 1950's. Generally the leaching mechanisms of these processes are not understood fundamentally. The reasons for this are the inherently complex nature of sulphide chemistry and that the sulphide concentrates usually consist of highly intergrown sulphide minerals. The leaching kinetics of sulphide concentrates where only one metal-sulphide mineral occurs have been investigated intensively, but not for sulphide concentrates with more than one metal-sulphide mineral. The behaviour of these mixed metal-sulphide minerals has mostly been investigated on plant scale to qualitatively determine the leaching trends of the process. The consequence of the relatively unKnown leaching mechanism and kinetics is that these processes are not controlled efficiently. This study was conducted on the acid-oxygen pressure leaching of Ni-Cu matte (the first stage leach process at the Ni-Cu refinery of Impala Platinum Ltd.). As a first step to improve the control efficiency of the process, the process must be stabilised. Therefore, an off line computer simulation program is proposed to control the repulping section of the plant that has previously been controlled solely by an operator. Controlling the repulping section is very important, because conditions exist in the repulping tanks for leaching to occur. This causes perturbations in the pulp entering the pressure leach autoclave. Due to the fast reaction kinetics of the matte in the pressure leach autoclave the perturbations entering the autoclave will influence the performance of the acid-oxygen pressure leach process. The simulation program was tested on the plant and indicated that considerable improvement in the stability of the operation could be achieved. In obtaining a better understanding of the behaviour of this process, it is essential that key variables and trends are identified. A methodology is proposed to analyse and model this ill-defined and poorly understood process from historical data by v artificial neural networks (ANN), inductive learning by decision trees and statistical techniques. The back propagation neural network, learning vector quantization neural network and the decision trees yielded comparable classification rates between 73% and 84%, and could serve as a basis for the adjustment of operating conditions to improve the efficiency of the process. The relative importance of the process variables is determined by a method of sensitivity analysis and together with the statistical mean, the effect of an increase or decrease in the variable on the process is quantified. These results are substantiated by experimental findings. A leaching mechanism for the acid-oxygen pressure leach of Ni-Cu matte is postulated. The leaching sequence of the nickel and copper sulphides is Ni3Sr Ni7S" NiS-Ni3S4, and CU2S-CU31SWCU1.BS-CUS, respectively. Ni7Sa and CU31 S1a are intermediate nickel and copper sulphide phases that form during the leaching process. Ni alloy has a galvanic effect on the sulphide minerals which inhibits the overall leaching rate and results in the formation of H2S and the intermediate nickel and copper sulphides (Ni7Sa and CU31 S1a). A semi-empirical kinetic model was developed based on the chemical reaction rate expressions of the leaching mechanism. This model can accurately simulate the batch leaching process for variations in the oxygen partial pressure, oxygen flowrate, temperature, particle size, initial acid concentration and pulp density. A sensitivity analysis on the model indicated that for a matte with a lower initial Ni alloy content the leaching rate of nickel is much faster.

Thesis (PhD (Process Engineering))--University of Stellenbosch, 2000. / Stricter government regulations and higher energy costs have forced the chemical industry to focus more on environmentally friendly processes and to reduce energy consumption. The main goals of chemical companies are to obtain a high product yield and selectivity, and to reduce unwanted side products. Furthermore, if ractions can be performed at lower temperature, while maintaining the reaction conversion, it will result in large energy savings. Low termperature dehydrogenation reactions (below 300 degrees C) are very selective and do not produce many by-products, but conversion is limited by the reaction equilibrium. The conversion limitations have resulted in the development of alternative processes in recent years for producing alkenes from alkanes and aldehydes or ketones from alcohols....

Waxes have many potential uses but large-scale application is hampered by their virtual insolubility. By grafting the wax with a polyethylene glycol segment to form an alcohol ethoxylate, the solubility of the wax in commercial solvents is significantly increased. Alcohol ethoxylates are produced by the polymerisation addition of ethylene oxide onto an oxidised wax. Current methods of alcohol ethoxylate production from alcohols lead to wide ethylene oxide addition distribution and large quantities of residual alcohol. The objective of this study is to provide a method for narrowing the ethylene oxide distribution and to reduce the residual alcohol content. It is proposed to concentrate the alcohol ethoxylate in a post-production separation process using supercritical fluid extraction. The system is modelled to contain three pseudo-components: an alkane, an alcohol and an alcohol ethoxylate. Propane is selected as the supercritical solvent of choice due to the large solubility difference between the alkane and polyethylene glycol. Lower molecular weight alkane phase equilibrium data with propane is abundant but extrapolation to higher molecular weights requires further investigation as it may be complicated by molecular folding. Molecular folding occurs in crystalline polyethylene and high molecular weight normal alkanes but information regarding molecular folding in solution is inconclusive. A model is proposed for molecular folding of normal alkanes in supercritical solution. A high molecular weight alkane mixture is synthesised and phase equilibrium measurement with propane are conducted. A lower molecular weight alkane mixture is used to prove the application of the principle of congruency to high-pressure phase equilibria. In the high wax mass fraction region the measurements are between the no-folding and once-folded relationship, indicating the possibility of partial molecular folding. In the mixture critical and low wax mass fraction region the measurements are similar to the non-folding relationship. Molecular folding in solution is thus dependent on the solution concentration. No phase equilibria measurements exist for propane with either high molecular weight alcohols or alcohol ethoxylates. Measurements of propane with an alcohol mixture show total solubility below 140barA for temperatures up to 408K. Measurements of propane with an alcohol ethoxylate at temperatures between 378 and 408K shows that for an alcohol ethoxylate mass fraction between 0.025 and 0.5 pressures greater than 275barA are required for solubilisation. When comparing the solubility of the three pseudo-components, the alkane is the most soluble followed by the alcohol. The alcohol ethoxylate is the least soluble. A counter-current supercritical extraction process is proposed for the concentration of the alcohol ethoxylate. Pilot plant tests were conducted and the proposed set-up shows good separation. An estimate of the energy requirements shows that heating and cooling constitute the majority of the energy required but with the use of heat integration it can be reduced by approximately 33%. This work thus shows that the proposed process is both technically and economically viable. Although this work has provided a method for concentrating the alcohol ethoxylate, the process has not been optimised yet and future work includes the fine-tuning of this process.

Thesis (PhD (Process Engineering))--University of Stellenbosch, 2007. / Nickel-Copper sulphide ores are the most important Platinum Group Metal bearing ores. The South African deposits are exceptionally rich in the platinum group metals (PGMs) and production of the PGMs is the primary purpose of treating these ores. The methods used in the recovery of the PGMs from the nickel-copper ores generally consists of ore concentration by physical techniques, pyrometallurgical concentration and hydrometallurgical extraction of the base metals followed by the PGMs. Pyrometallurgical concentration produces Ni-Cu matte, which is treated by hydrometallurgical processes to recover the nickel, copper, cobalt and the precious metals. In this study, the leaching behaviour of a Ni–Cu matte in CuSO4–H2SO4 solution during the repulping (pre-leach) stage at Impala Platinum Refineries was studied. The repulping stage is basically a non–oxidative atmospheric leach stage, in which nickel, iron and cobalt are partially dissolved, while the copper is precipitated. To understand the nature of the leaching process during this stage of the base metal refining operation, the effects of variations in the key process variables such as temperature, stirring rate, particle size, pulp density, residence time, initial copper and acid concentrations were investigated. The pre-leached matte was then pressure leached to ascertain the effect of process conditions in the pre-leach stage on the subsequent pressure leach stage. It was found that the leaching mechanism entails dissolution of metal alloys out of sulphide minerals with transformation of Ni3S2 to NiS. Aqueous copper precipitates as metallic copper and as chalcocite. The matte is leached by both acid and the cementation process, especially in the early stage when the Cu2+ ions are present. Galvanic interaction of the sulphide minerals and/or the Ni alloy also enhances the leaching process. The leaching kinetics of Ni was characterized with the shrinking core model and was found to be controlled by diffusion through surface layer. An activation energy of 31 kJ/mol was obtained, which also suggested a diffusion controlled leaching reaction. Atmospheric leaching tests indicated that Ni extraction increased slightly in the temperature range 50 – 60 oC, however no significant increase was observed from 60 to 80 oC, probably because the leaching process was found to be diffusion controlled. The slight increase in nickel dissolution at higher temperatures (>60 oC) may be attributed to the transformation of Ni3S2 to NiS, which is easier to leach. Co extraction appeared to be insensitive to temperature changes, while Fe extraction was low at 50 oC but increased significantly at 60 – 80 oC. The Ni extraction increased gradually with increase in the stirring rate from 145 to 400 rpm while Co and Fe extractions were insensitive at 145 and 205 rpm, but increased substantially at 400 rpm. This was probably due to increased mass transfer rate and transformation of Ni3S2 to NiS. With pulp density, Ni and Co extractions appeared to be insensitive to changes in the pulp density as only a slight increase in extractions was observed when the density was reduced from 1.7 kg/L to 1.6 kg/L. Similar iron extractions were achieved at 1.7 and 1.75 kg/L but increased significantly at 1.6 kg/L. It was found that Ni and Co extractions were not significantly affected by changes in the particle size, probably because metal alloys were liberated and hence exposed to the leaching solution. Iron extraction could not be determined accurately because of iron precipitation at pH above 3. Generally the leaching of metals did not depend on the initial copper concentration in the investigated range of 25 - 48 g/L Cu. It was also observed that initial acid concentration did not have an effect on Ni extraction, probably due to the fact that most of the nickel was leached by the process of cementation. However, Co and Fe extractions increased when the acid increased from 90 g/L to 110 g/L, but no further increase was noted at 125 g/L. A residence time of 5 hours was found to be adequate as there was no significant increase in metal extractions when the residence time was increased beyond 5 hours. As much as 20% Ni, 40% Co and 80% Fe can be extracted from the Ni-Cu matte during the repulping stage of the leaching process studied, provided the investigated conditions prevail. Generally the rate of Cu cementation increased with increasing temperature and pulp density, but decreased with an increase in particle size, acid and copper concentrations. The rate of stirring did not affected Cu cementation. It was found that aqueous copper precipitated from the solution within 90 minutes when the temperature was raised to 80 oC. However, under the present pre-leach temperature of about 60 oC complete Cu cementation can only be achieved after about 5 hours. The cementation reaction was found to follow a mixed control mechanism, with two distinct activation energies namely 18.2 kJ/mol at 70 – 80 oC and 74.6 kJ/mol at 50 – 70 oC. This suggested that the rate of cementation reaction is probably controlled by a boundary layer diffusion mechanism at higher temperatures. At low temperatures the rate is probably controlled by a surface reaction mechanism. The pressure leaching experiments, which were aimed at investigating the response of pre-leached matte to the subsequent pressure leaching process, showed that Ni extractions were similar for the investigated pre-leach temperature of 50 – 80 oC and stirring rate of 145 – 400 rpm. For the pre-leach stage conditions of 60 – 80 oC and 205 – 400 rpm Ni3S2 was transformed into NiS, which is easier to leach in the pressure leaching stage. However, because of the aggressive conditions prevailing in the pressure leaching stage, all the nickel minerals were leached at about the same rate. In the case of pulp density, Ni extraction was comparable for all the investigated pulp densities (1.6 – 1.75 kg/L). This was probably due to the fact that Ni3S2 transformed to NiS in the pre-leach stage. It was found that Ni extraction increased with increasing residence time for the investigated time of 1 hour to 9 hours, probably due to the changes in the mineral phases of the matte as indicated above. The copper minerals (Cu2S and Cu1.96S) transformed into Cu2S and Cu1.8S with aqueous copper being precipitated, and were not leached under the applied conditions. All the cobalt and iron dissolved in the pressure leaching stage. A semi-empirical kinetic model was developed for the pre-leaching stage. A comparison of the model predictions and the experimental data for the dissolved species during the batch leaching process showed that the model can satisfactorily fit the trends in the leaching of the metals.

Thesis (PhD)--University of Stellenbosch, 2004. / 388 leaves printed on single pages, preliminary pages i- xv and numbered pages 1-371. Includes bibliography, list of tables and figures. / Digitized at 330 dpi black and white and 330 dpi color PDF format (OCR), using KODAK i 1220 PLUS scanner. / ENGLISH ABSTRACT: Bath type furnaces have become an established technology for the intensive smelting, converting and refining of primary and secondary raw materials. Since these furnaces normally have large inventories, long time constants and complex metallurgies, a dynamic model-based prediction strategy is the only feasible approach to operator decision support and process control. This dissertation presents a semi-empirical approach to the stochastic modelling of bath-type pyrometallurgical reactors, which leads to a generic model type called the Equilib-ARMAX model. The modelling approach is applied to three case studies: • A nickel-copper matte converting operation using a submerged lance injection reactor • A chromite smelting operation to produce high carbon ferrochrome using a direct current (DC) plasma smelting furnace • An ilmenite smelting operation to produce high titania slag and pig iron, using a direct current (DC) plasma smelting furnace. In each case, the industrial operations were analysed with regard to the practical and technological constraints which influence the type and quality of the process data. The fundamental process phenomena associated with each operation have been analysed to ascertain which fundamental variables should be included within the overall semi-empirical approach, without sacrificing model transparency, simplicity, accuracy and calculation time. It was considered that an overly complex model would be inappropriate given that data from industrial smelting operations show significant random variance. The thermochemistry and phase equilibria associated with each operation are discussed in detail, as they become the fundamental backbone of the semi-empirical models. The equilibria have been modelled with software that uses non-ideal solutions models and Gibbs free energy minimisation to predict the phase and chemical equilibria that could be expected for a given feed recipe and operating temperature. As the thermodynamic modelling software is not stable within an industrial environment, an artificial intelligent mapping technique has been developed to map process inputs to equilibrium outputs. A multi-layer perceptron neural network has been used as the convenient mapping method to represent equilibrium. The neural networks were trained using tens of thousands of feed recipes, where the feed component ratios were varied based on a 3N factorial design. The amounts and chemistries of all equilibrium phases could be calculated with high accuracies (R2 > 0.95) in all cases. Further stochastic analysis and modelling require additional information about the property distributions associated with each measurement. The homogeneities of the furnace products (slag, alloy and flue dust) critically influence the level of confidence that one can associate with plant measurements. The homogeneities were characterised for the DC plasma arc furnaces and they were benchmarked against a submerged arc furnace. It was found that the homogeneity varied per element, with silicon and sulphur tending to show highest variations in the alloy melts. The observation that the variation in these two elements are both high can partially be attributed to the fact that SiS evaporates from the bath surface, especially in regions close to the arc attachment zone. A significant negative correlation was found between the relative standard deviation per tap (using silicon) and the degree of superheat / subcooling of the alloy, indicating that the homogeneity can be strongly influenced by the changes in rheology due to subcooling below the liquidus (which leads to the precipitation of solid phases and increases the observed melt viscosity). Mixedness or homogeneity and data uncertainty are therefore inseparably linked. The relative standard deviations associated with the homogeneity characterisation, as well as known sampling and assaying variances were used to develop reconciled material balances based on measured plant data. Material balance closure was therefore obtained within the inherent uncertainties of the plant data. Biases in the plant data were identified simultaneously with data reconciliation. Moreover, it was shown using Fast Fourier Power Spectra and statespace analysis that the data reconciliation was a good low-pass filter, as it extracted the major process trends components in the noisy data and it also improved the overall dynamic behaviour characteristics of the data. Finally systems identification techniques were used to develop dynamic transfer function models that were linear in the parameters to be estimated. These systems models were based on the reconciled plant data and equilibrium predictions. The final systems models are therefore equilibrium-autoregressive-moving-average models with exogenous variables (Equilib-ARMAX). The model parameters can be estimated recursively using a simple least squares method. The final models could dynamically predict the metallurgy of the subsequent tap 4-6 hours in advance, based on a given suite of set-points, within the inherent accuracy of the data. These models may be used to suggest the optimal operating conditions through an operator guidance system, or more simply, the models are simple enough to be used in a spreadsheet on a manager's desk. / AFRIKAANSE OPSOMMING: Bad-tipe oonde is reeds 'n gevestigde tegnologie wat algemeen gebruik word vir die intensiewe smelting, omsetting en raffinering van primere en sekondere roumateriale. Aangesien hierdie oonde normaalweg groot inventarisse, lang tydkonstantes en komplekse metallurgiee het, is dinamiese, modelgebaseerde voorspelling die enigste uitvoerbare benadering tot operateur besluitnemingsteunstelsels en prosesbeheer. Hierdie proefskrif stel 'n nuwe generiese, semi-empiriese benadering voor om die bad-tipe oonde stogasties te modelleer en lei tot die sogenaamde Equilib-ARMAX model. Die modelleringsbenadering word geevalueer deur drie gevallestudies: • 'n Nikkel-koper swawelsteen omsettingsproses in 'n dompel-Ians inspuit reaktor • 'n Chromiet smeltingsproses om hoe-koolstof ferrochroom te produseer in 'n gelykstroom (GS) plasmaboogoond • 'n Ilmeniet smeltingsproses om hoe titania slak en ruyster te produseer in 'n gelykstroom (GS) plasmaboogoond. In elke geval is die industriele prosesse ontleed met betrekking tot die praktiese en tegnologiese beperkings wat die tipe en die gehalte van die prosesdata beinvloed. Die fundamentele prosesgedrag van elke proses is ontleed om te bepaal welke fundamentele veranderlikes ingesluit moet word in die semi-empiriese benadering, sonder om model deursigtigheid, eenvoud, akkuraatheid en berekeningstyd in te boet. Die ontwikkeling van oor-komplekse modelle is beskou as ongepas, gegewe dat die data van industriele smeltingsprosesse beduidende onsekerhede toon. Die termochemiese en fase-ewewigte geassosieer met elke proses word breedvoerig bespreek, aangesien dit die fundamente1e grondslag van die semi-empiriese modelle verskaf. Die ewewigte is gemodelleer met rekenaar simulasie-programmatuur wat nie-ideale oplossingsmodelle en Gibbs vrye-energie minimering gebruik om die fase en chemiese ewewigte, wat verwag kan word vir 'n gegewe toevoerresep en bedryfstemperatuur, te voorspel. Aangesien termodinamiese modelleringsprogrammatuur normaalweg nie stabiele gedrag toon in 'n intydse industriele omgewing nie, word kunsmatig intelligente projeksietegnieke gebruik om prosesinsette te projekteer na die ekwavilente ewewigsvoorspellings. 'n Multilaag perseptron neurale netwerk is gebruik as 'n eenvoudige metode om hierdie ewewigsprojeksies voor te stel. Die neurale netwerke is afgerig deur van tienduisende toevoer resepte gebruik te maak. Die verhoudings van die komponente in die voer is gewissel gebaseer op 'n 3N faktoriaalontwerp. Die hoeveelhede en samestelling van al die ewewigsfases kon in alle gevalle bereken word met hoe akkuraatheid (R2 > 0.95). Verdere stogastiese analise en modellering is slegs moontlik met kennis oor die eienskapsverspreidings geassosieer met elke komponent. Die homogeniteite van die oondprodukte (slak, legering en vlieg-as) bepaal, tot 'n groot mate, die betroubaarheidsvlak van die aanlegmetings. Homogeniteite is gekarakteriseer vir die GS-plasmaboogoonde en is vergelyk met die homogeniteite wat in dompelboogoonde gevind word. Die homogeniteite het gevarieer per komponent. Silikon en swawel neig om die grootste ruimtelike variasies te toon in die legerings wat bestudeer is. 'n Beduidende negatiewe korrelasie is gevind tussen die relatiewe standaardafwyking per tap (gebaseer op silikon) en die graad van superverhitting / onderverkoeling van die legering. Dit dui aan dat die homogeniteit sterk beinvloed word deur veranderinge in die smelt reologie. Vermenging, reologie, homogeniteit en data onsekerheid (integriteit) is daarom ten nouste gekoppel. Die relatiewe standaardafwykings geassosieer met die homogeniteitsbepaling, asook die monsternemings- en ontledingsvariansies, is gebruik om die aanlegdata te rekonsilieer onderhewig aan die behoud van die komponent en totale stroom massabalanse. Die massabalanse is dus gesluit deur aanpassings aan die metings te maak binne die inherente onsekerhede in die data. Sistematiese foute in die data is gelyktydig met die rekonsiliasie geidentifiseer. Verder is deur diskrete Fourier energiespektra en toestand-ruimte analises getoon dat massabalans-rekonsiliasie dien as 'n goeie seinfilter om hoe-frekwensie geraas te verminder en tergelykertyd die dinamiese gedragseienskappe van die data te verbeter. Stelsel-identifikasietegnieke is gebruik om dinamiese oordragsfunksiemodelle te ontwikkel wat linieer is met betrekking tot die modelparameters. Hierdie stelselmodelle is gebaseer op gerekonsilieerde data, eksogene prosesdata en ewewigsberekeninge, en word vervolgens ewewigs-autoregressiewe-lopende-gemiddelde modelle met eksogene veranderlikes (Equilib ARMAX) genoem. Die modelparameters kan deur gewone kleinste-kwadrate metodes beraam word. Die finale modelle kan die metallurgie van toekomstige tappe 4-6 uur voortydig voorspel, gebaseer op beskikbare stelpunte en binne die inherente presisie van die data. Hierdie modelle kan gebruik word om optimale bedryfskondisies vir prosesbeheer te identifiseer, en is eenvoudig genoeg om in sigbladformaat op 'n aanlegbetuurder se rekenaar gebruik te kan word.

Thesis (PhD)--University of Stellenbosch, 2004. / ENGLISH ABSTRACT: In this study an equation of state has been developed for the specific purpose of representing systems of simple non-polar spherical and chain-like components and their mixtures for practical applications. To be applied in engineering calculations, the model has to be accurate, be able to represent mixtures with large size asymmetry without the use large binary interaction parameters, and be mathematically simple enough to ensure rapid computations. The model is developed through a sequential evaluation of the statistical mechanical theory of particles and the various approaches available to extend it to real fluid systems. The equation of state developed in this work models the real fluid systems as interacting with a highly simplified two step potential model. The repulsive interactions are represented by a newly developed simplified form of the hard sphere equation of state, capable of representing the known hard sphere virial coefficients and phase behaviour to a high degree of accuracy. This equation has a realistic closest packed limiting density in between the idealised hard sphere fluid random and crystal structure limits. The attractive interactions between the particles are incorporated into the model through a perturbation expansion represented in the form of a double summation perturbation approximation. The perturbation matrix was optimised to have the lowest order in density necessary to still be able to accurately represent real fluid properties. In a novel approach to obtain simple mixing rules that result in the theoretically correct second virial coefficient composition dependence, the perturbation matrix is constrained in such a manner that only the first perturbation term has a term that is first order in density. From a detailed evaluation of the various methods available to represent chain-like non-spherical systems it was finally concluded that the Perturbed Hard Chain Theory provided an ideal compromise between model simplicity and accuracy, and this method is used to extend the equation to chain-like systems. Finally the model is extended to fluid mixtures by uniquely developed mixing rules resulting in the correct mixture composition dependence both at low and high system densities. The newly developed equation of state is shown to be capable of representing the pure component systems to a comparable degree of accuracy as the generally applied equations of state for non-spherical systems found in the literature. The proposed equation is furthermore also shown equal or improve on the predictive ability of these models in the representation offluid mixtures consisting out of similar chainlike or size and energetic asymmetric components. Finally, the computational time required to model the behaviour of large multi-component fluid mixtures using the new equation of state is significantly shorter that that of the other semi-empirical equations of state currently available in the literature. / AFRIKAANSE OPSOMMING: Hierdie werkstuk behels die ontwikkeling van ‘n toestandsvergelyking wat spesifiek gerig is op toepassings in alledaagse, praktiese ingenieurstipe berekeninge en daartoe instaat is om sisteme bestaande uit nie-polêre spferiese- en ketting-tipe komponente en hulle mengsels teKettingteorie (PHCT) die mees geskikde metode is vir hierdie doel en is op die vergelyking toegepas. As ‘n laaste stap in die toestandsvergelykingontwikkelling is daar mengreëls ontwikkel vir die vergelyking wat die korrekte samestellingsafhanklikheid toon vir beide die lae en hoë digtheidskondisies. Die model wat in hierdie studie ontwikkel is, is met verskeie ander bekende toestandsvergelykings, wat daartoe instaat is om nie-spferiese sisteme te modelleer, vergelyk en daar is gevind dat die nuwe model daartoe instaat is om suiwer sisteme net so goed as die bestaande vergelykings te modelleer. Verder is daar ook gevind dat die nuwe vergelyking die modellering van verskeie mensels van kettingtipe komponente en komponente van uiteenlopende groottes of interaksie energieë kan ewenaar of verbeter. Laastens is daar ook gevind dat die tyd nodig vir die modellering van die termodinamiese gedrag van mengsels van ‘n groot hoeveelheid komponente aansienlik korter is vir die nuwe model as die ander bekende semi-empiriese vergelykings. kan beskyf. Om aan hierdie vereistes te voldoen moet die toestandsvergelyking die relevante sisteme akkuraat kan modelleer, slegs klein interaksie parameters benodig om mengsels van komponente met groot verskille in molekulêre groottes akkuraat voor te stel en steeds wiskundig eenvoudig genoeg wees om vinnige berekeninge te verseker. Die vergelyking is ontwikkel deur ‘n sistematiese evaluering van die statisitiese meganiese teorie van partikels en die verskillende metodes om hierdie teorië op werklike sisteme toe te pas. Die toestandsvergelyking beskryf die intermolekulêre interaksie tussen die verskillende komponente met ‘n hoogs vereenvoudigde twee-stap interaksie potensiaal model. Die afstotende kragte tussen die komponente word in ag geneem deur ‘n nuwe vergelyking wat ontwikkel is om die gedrag van ‘n ideale harde spfeer sisteem te modelleer. Hierdie hardespfeermodel is daartoe instaat om die viriale koeffisiënte en die fase gedrag van teoretiese harde spfeer sisteme akkuraat te modelleer, en het ‘n maksimum digtheidslimiet wat tussen teoretiese waardes van ‘n perfek geordende en nie-geordende harde spheer sisteem lê. Die aantrekkinskragte tussen die partikels word beskou as ‘n perturbasie van die harde-spheer vergelyking. ‘n Term bestaande uit ‘n dubbelle sommasiefunksie word gebruik om hierdie perturbasie uitbreiding voor te stel. Die sommasie term is geoptimiseer sodat die finale toestandsvergelyking die laagste digtheidsgraad het wat steeds tot ‘n akkurate voorstelling van die termodinamiese gedrag van werklike sisteme lei. Die sommasiefunksie is so gespesifiseer dat die eerste term van die perturbasie uitbreiding slegs ‘n eerste graadse orde in digtheid het in ‘n unieke benadering om te verseker dat die mengreëls van die toestandsvergelyking die teoreties korrekte samestellingafhanklikheid van die mengselvirialekoeffisiente tot gevolg het. ‘n Deeglike ondersoek van die verskillende metodes om die toepassing van die toestandsvergelyking uit te brei tot die moddellering van nie-spheriese ketting-tipe molekules is gedoen en daar is uiteindelik tot die gevolgtrekking gekom dat die Geperturbeerde Harde