Production of platinum and associated metals is a major source of revenue for South Africa. Significant losses occur in the concentrating stage (10 to 15 per cent) and this research is focused on optimising platinum flotation. Research begins by conducting laboratory batch flotation tests. However, subsequent pilot-plant tests often produce different results. It is believed these differences arise from the artificial nature of laboratory techniques. This project was focused on improving flotation techniques in the laboratory. The largest source of platinum in South Africa is the UG2 reef and two samples of this ore were used for testing: ‘good’ and ‘bad’ ore. These had different characteristics with regards to the recovery of PGMs and the presence of talc. The latter is an unwanted floatable mineral, which must be depressed to prevent excessive recovery. The conventional laboratory test procedure makes use of batch tests in various sizes of flotation cells. The procedure was made more realistic, by using four stages of flotation, rather than just two, to mimic a typical platinum flotation plant. The use of four stages made it possible to separate the fast-floating and slow-floating stages and to control froth conditions accordingly. Attention was also given to the fact that in laboratory tests, water is often added to the ‘cleaner’ stage of flotation, to make up the level. Experiments showed that this dilution, which does not take place in practice, had a significant impact on overall efficiency. A method of measuring frother concentration was developed and used to determine the realistic level of frother in cleaning tests. Tests at these levels of frother concentration showed that significant improvements could be made to plant performance, by making use of a thickener to reduce the frother concentration in the cleaning stages. The improved test procedure was used on both good and bad ores, and the effect of regrinding was also tested. A combined solids recovery of 2 % over both cleaners was targeted for all test work. At this recovery, the regrinding of the bad ore increased the PGM recovery from 67 to 76 per cent at the cost of an additional 8 g/t depressant. An investigation of the effect of frother concentration in the cleaning stage, using good ore, demonstrated that that rejection of chromite could be improved significantly by reducing frother concentration. The tests mimicked the use of a thickener to separate some of the water with a high concentration of frother. Tests conducted on the good ore showed that use of two thickeners, as opposed to none, reduced the Cr2O3 content of the final concentrate from 4.2 to 3.2 per cent for the equivalent concentrate mass and PGM recovery. The depressant requirement was also reduced from 67 to 55 g/t. These tests provided insight on how to improve performance on a platinum flotation plant, particularly when floating the bad ore.

Abstract available in PDF copy. / Thesis (M.Sc.-Chemical Engineering)-University of Natal, 1988.

The Fischer-Tropsch reaction technology is utilised in Sasol’s Coal-to-liquid plant to produce liquid fuels from low grade coal. There are several processes on the Coal-to-liquid plant that generate aqueous streams which contain a high organic load and require treatment. The main contributors to the wastewater are the Phenosolvan plant, producing stripped gas liquor (SGL), and the Synthol plant, producing Fischer-Tropsch reaction water (FTRW). Stripped gas liquor contains water, organic acids, ammonia, and potentially toxic phenols. Fischer-Tropsch reaction water contains volatile fatty acids and alcohol. Stripped gas liquor is therefore nitrogen-rich while FTRW is nitrogen-deficient and requires nutrient supplementation for anaerobic treatment. Therefore co-treatment of the two streams could reduce nitrogen supplementation requirements. This study is part of a larger project to determine the feasibility of anaerobically co-digesting FTRW and SGL. This study has looked at the influence of SGL on the methanogenic activity of FTRW-acclimated sludge and involved the development of a method which allows accurate recording of the methanogenic activity in batch assays. Other studies involving the anaerobic digestion of high phenolic wastewaters showed that the phenol had an inhibitory effect on the specific methanogenic activity of the sludge, which was not acclimated to the phenol. The objective of this work was to test the hypotheses that (1) anaerobic sludge acclimated to FTRW will be inhibited by high molecular weight organics in SGL and (2) FTRW-acclimated sludge will not degrade phenolic compounds in SGL. This information will be used for designing process configurations for simultaneous treatment of the two streams with minimum contamination of the effluent stream. The serum bottle was used as a small batch reactor and the biogas production was monitored as an indication of the state of the reaction. The biogas produced was collected and measured by the downward displacement of a sodium hydroxide solution, which absorbed the carbon dioxide and collected only the methane. A concentration of 1 g COD/ℓ FTRW was chosen as the reference test due to the reproducibility of the replicates within each experiment as well as its reproducibility across different batches of sludge. For the first inhibition test, the test units contained an additional 5% SGL (0.05 g COD/ℓ SGL) and an additional 15% SGL (0.15 g COD/ℓ SGL, i.e. 13% of the total COD load) respectively, added to 1 g COD/ℓ FTRW. The 5% SGL test unit showed no inhibition compared to the reference unit. There was a reduction in the specific methanogenic activity of the 15% SGL test units compared to the reference unit. Since the total COD load was not the same in each unit, it cannot be conclusively stated that the SGL was responsible for the reduction in SMA, but this seems a reasonable possibility in the light of results from the reference test selection experiments which showed higher SMA at higher organic loading rates. For the second inhibition test, the test units contained 85% FTRW (0.85 g COD/ℓ FTRW) and 15% SGL (0.15 g COD/ℓ SGL) to make up a total COD load of 1 g COD/ℓ. There was an increase in the specific methanogenic activity of the test unit compared to the reference unit. There was very little change in the phenol concentration. Therefore, it was concluded the addition of SGL potentially reduced the SMA and that this could be an inhibitory effect, but that any inhibition would be a function of the concentration of potentially inhibitory substances in SGL and that these concentrations vary from batch to batch. However, the degree of SMA reduction is fairly low and would not prevent co-digestion of the two streams at the concentrations tested. It has been shown that FTRW anaerobic digestion can proceed adequately in the presence of SGL. There was some evidence that phenolics were degraded but at a much slower rate than COD. The percentage reduction in SMA due to additional SGL at concentrations and SGL:FTRW ratios tested was between 0 and 51%. Ultimately, this work is a first step in the development of a co-digestion model relating organic loading rate, SGL:FTRW feed ratio to methane recovery and extent of biodegradation of phenol for use in the design and optimization of a co-digestion system. / M.Sc.Eng. University of KwaZulu-Natal, Durban 2014.

The objective of this study was to provide a technique, based on the residence time distribution of a process, for modelling, assessing and improving flow in the processes of water and waste water treatment works. The technique should be accessible to the staff managing and operating the works. From a review of the literature, a preference was given for the experimental method used for determination of the tracer response, including choice of tracer and tracer addition and monitoring. Data analysis techniques were reviewed, and the method of time domain fitting was developed into a computer program, IMPULSE. IMPULSE provided a tool for analysis of residence time data, and removed the constraint of numerical complexity. Using the building blocks of IMPULSE, a realistic flow model can be constructed from tracer data and evaluated. IMPULSE allows a quantitative comparison of models proposed for a process, and provides the parameters of the models. These parameters quantify the non-idealities in a process. A knowledge of the non-idealities provides a basis for decision-making when modifying a process. The results of tracer experiments performed on some water and waste water treatment processes were analysed using IMPULSE. The results showed that collection of good experimental data was critical to the success of the analysis. It is proposed that a guide be produced which draws out the main points raised in the study, including collection of tracer data and use of IMPULSE. The guide should be accessible and easily understandable to the staff managing and operating water and waste water treatment works. / Thesis (M.Sc.Eng.)-University of Natal, Durban, 1995.

Thesis (PhD)--University of Stellenbosch, 2004. / 271 Leaves printed single pages, preliminary pages i-xviii and 253 numberd pages. Includes bibliography. List of figures, List of tables. / ENGLISH ABSTRACT: A Neurocontrol Paradigm for Intelligent Process Control using Evolutionary Reinforcement Learning Balancing multiple business and operational objectives within a comprehensive control strategy is a complex configuration task. Non-linearities and complex multiple process interactions combine as formidable cause-effect interrelationships. A clear understanding of these relationships is often instrumental to meeting the process control objectives. However, such control system configurations are generally conceived in a qualitative manner and with pronounced reliance on past effective configurations (Foss, 1973). Thirty years after Foss' critique, control system configuration remains a largely heuristic affair. Biological methods of processing information are fundamentally different from the methods used in conventional control techniques. Biological neural mechanisms (i.e., intelligent systems) are based on partial models, largely devoid of the system's underlying natural laws. Neural control strategies are carried out without a pure mathematical formulation of the task or the environment. Rather, biological systems rely on knowledge of cause-effect interactions, creating robust control strategies from ill-defined dynamic systems. Dynamic modelling may be either phenomenological or empirical. Phenomenological models are derived from first principles and typically consist of algebraic and differential equations. First principles modelling is both time consuming and expensive. Vast data warehouses of historical plant data make empirical modelling attractive. Singular spectrum analysis (SSA) is a rapid model development technique for identifying dominant state variables from historical plant time series data. Since time series data invariably covers a limited region of the state space, SSA models are almost necessarily partial models. Interpreting and learning causal relationships from dynamic models requires sufficient feedback of the environment's state. Systemisation of the learning task is imperative. Reinforcement learning is a computational approach to understanding and automating goal-directed learning. This thesis aimed to establish a neurocontrol paradigm for non-linear, high dimensional processes within an evolutionary reinforcement learning (ERL) framework. Symbiotic memetic neuro-evolution (SMNE) is an ERL algorithm developed for global tuning of neurocontroller weights. SMNE is comprised of a symbiotic evolutionary algorithm and local particle swarm optimisation. Implicit fitness sharing ensures a global search and the synergy between global and local search speeds convergence.Several simulation studies have been undertaken, viz. a highly non-linear bioreactor, a rigorous ball mill grinding circuit and the Tennessee Eastman control challenge. Pseudo-empirical modelling of an industrial fed-batch fermentation shows the application of SSA for developing partial models. Using SSA, state estimation is forthcoming without resorting to fundamental models. A dynamic model of a multieffect batch distillation (MEBAD) pilot plant was fashioned using SSA. Thereafter, SMNE developed a neurocontroller for on-line implementation using the SSA model of the MEBAD pilot plant. Both simulated and experimental studies confirmed the robust performance of ERL neurocontrollers. Coordinated flow sheet design, steady state optimisation and nonlinear controller development encompass a comprehensive methodology. Effective selection of controlled variables and pairing of process and manipulated variables were implicit to the SMNE methodology. High economic performance was attained in highly non-linear regions of the state space. SMNE imparted significant generalisation in the face of process uncertainty. Nevertheless, changing process conditions may necessitate neurocontroller adaptation. Adaptive neural swarming (ANS) allows for adaptation to drifting process conditions and tracking of the economic optimum online. Additionally, SMNE allows for control strategy design beyond single unit operations. SMNE is equally applicable to processes with high dimensionality, developing plant-wide control strategies. Many of the difficulties in conventional plant-wide control may be circumvented in the biologically motivated approach of the SMNE algorithm. Future work will focus on refinements to both SMNE and SSA. SMNE and SSA thus offer a non-heuristic, quantitative approach that requires minimal engineering judgement or knowledge, making the methodology free of subjective design input. Evolutionary reinforcement learning offers significant advantages for developing high performance control strategies for the chemical, mineral and metallurgical industries. Symbiotic memetic neuro-evolution (SMNE), adaptive neural swarming (ANS) and singular spectrum analysis (SSA) present a response to Foss' critique. / AFRIKAANSE OPSOMMING: 'n Neurobeheer paradigma vir intelligente prosesbeheer deur die gebruik van evolusionêre versterkingsleer Dit is 'n komplekse ontwikkelingstaak om menigte besigheids- en operasionele doelwitte in 'n omvattende beheerstrategie te vereenselwig. Nie-lineêriteite en vele komplekse prosesinteraksies kombineer om ingewikkelde aksie-reaksie verwantskappe te vorm. Dit is dikwels noodsaaklik om hierdie interaksies omvattend te verstaan, voordat prosesbeheer doelwitte doeltreffend gedoen kan word. Tog word sulke beheerstelsels dikwels saamgestel op grond van kwalitatiewe kriteria en word ook dikwels staatgemaak op historiese benaderings wat voorheen effektief was (Foss, 1973). Dertig jaar na Foss se kritiek, bly prosesbeheerstelsel ontwerp 'n heuristiese saak. Die biologiese prosessering van informasie is fundamenteel verskillend van metodes wat gebruik word in konvensionele beheertegnieke. Biologiese neurale meganismes (d.w.s., intelligente stelsels) word gebaseer op gedeeltelike modelle, wat grotendeels verwyderd is van die onderskrywende natuurwette. Neurobeheerstrategieë word toegepas sonder suiwer wiskundige formulering van die taak of die omgewing. Biologiese stelsels maak eerder staat op kennis van aksie-reaksie verhoudings en skep robuuste beheerstrategieë van swak gedefineerde dinamiese stelsels. Dinamiese modelle is of fundamenteel of empiries. Fundamentele modelle word ontwikkel vanaf eerste beginsels en word tipies uit algebraïese en differensiële vergelykings saamgestel. Modellering vanaf eerste beginsels is beide tydrowend en duur. Groot databasisse van historiese aanlegdata maak empiriese modellering aantreklik. Singuliere spektrumanalise (SSA) maak die vinnige ontwerp van empiriese modelle moontlik, waardeur dominante veranderlikes vanaf historiese tydreekse onttrek kan word. Aangesien tydreeksdata slegs 'n gedeelte van die prosesomgewing verteenwoordig, is SSA modelle noodwendig gedeeltelike modelle. Die interpretasie en aanleer van kousale verhoudings vanaf dinamiese modelle vereis voldoende terugvoer van omgewingstoestande. Die leertaak moet sistematies uitgevoer word. Versterkingsleer is 'n ramingsbenadering tot 'n doelwit-gedrewe leerproses. Hierdie tesis bewerkstellig 'n neurobeheerparadigme vir nie-lineêre prosesse met hoë dimensies binne 'n evolusionêre versterkingsleer (EVL) raamwerk. Simbiotiese, memetiese neuro-evolusie (SMNE) is 'n EVL algoritme wat ontwikkel is vir globale verstelling van die gewigte van ‘n neurobeheerder. SMNE is saamgestel uit 'n simbiotiese evolusionêre algoritme en 'n lokale partikelswerm-algoritme. Implisiete fiksheidsdeling verseker 'n globale soektog en die sinergie tussen globale en lokale soektogte bespoedig konvergensie.Verskeie simulasie studies is onderneem, o.a. die van 'n hoogs nie-lineêre bioreaktor, 'n balmeulaanleg en die Tennessee Eastman beheer probleem. Empiriese modellering van 'n industriële enkelladingsfermentasie demonstreer die aanwending van SSA vir die ontwikkeling van gedeeltelike modelle. SSA benader die toestand van 'n dinamiese stelsel sonder die aanwending van fundamentele modellering. 'n Dinamiese model van 'n multi-effek-enkelladingsdistillasie (MEBAD) proefaanleg is bewerkstellig deur die gebruik van SSA. Daarna is SMNE gebruik om 'n neurobeheerder te skep vanaf die SSA model vir die beheer van die MEBAD proefaanleg. Beide simulasie en eksperimentele studies het die robuuste aanwending van EVL neurobeheerders bevestig. Die gekoördineerde ontwerp van vloeidiagramme, gestadigde toestand-optimering en nie-lineêre beheerderontwikkeling vereis 'n omvattende metodologie. Beheerveranderlikes en die koppeling van proses- en uitvoerveranderlikes is implisiet en effektief. Maksimale ekonomiese aanwins was moontlik in hoogs nie-lineêre dele van die toestandsruimte. SMNE het besondere veralgemening toegevoeg tot neurobeheerderstrategieë ten spyte van prosesonsekerhede. Nietemin, veranderende prosestoestande mag neurobeheerderaanpassing genoodsaak. Aanpasbare neurale swerm (ANS) algoritmes pas neurobeheerders aan tydens veranderende proseskondisies en volg die ekonomiese optimum, terwyl die beheerder die proses beheer. SMNE bewerkstellig ook die ontwikkeling van beheerstrategieë vir prosesse met meer as een eenheidsoperasie. SMNE skaal na prosesse met hoë dimensionaliteit vir die ontwikkeling van aanlegwye beheerstrategieë. Talle kwelvrae in konvensionele aanleg-wye prosesbeheer word deur die biologies gemotiveerde benadering van die SMNE algoritme uit die weg geruim. Toekomstige werk sal fokus op die verfyning van beide SMNE en SSA. SMNE en SSA bied 'n nie-heuristiese, kwantitatiewe benadering wat minimale ingenieurskennis of oordeel vereis. Die metodologie is dus vry van subjektiewe ontwerpsoordeel. Evolusionêre versterkingsleer bied talle voordele vir 'n ontwikkeling van effektiewe beheerstrategieë vir die chemiese, mineraal en metallurgiese industrieë. Simbiotiese memetiese neuro-evolusie (SMNE), aanpasbare neurale swerm metodes (ANS) en singulêre spektrum analise (SSA) gee antwoord op Foss se kritiek.

Mineral processing plan ts, such as LONMIN's Eastern Platinum B-stream, typically have few on-line measurements, and key measures of performance such as grade only become available after samples have been analysed in the laboratory. More immediate feedback from a dynamic observer model promises enhanced understanding of the process, and facilitates prompt corrective actions, whether in open or closed loop . Such plant s easily enter sub-optimal modes such as large , uselessly re-circulating loads as the feed conditions change. Interpretation of such modes from key combinations of the variables deduced by an observer model , using a type of expert system, would add another level of intelligence to benefit operation. The aim of this thesis was to develop and implement a dynamic observer model of the LONMIN Eastern Platinum B-Stream into one of the existing control platforms available at the plant , known as PlantStar®, developed by MINTEK. The solution of the system of differential and algebraic equations resulting from this type of flowsheet modelling is based on an extended Kalman filter, which is able to dynamically reconcile any measurements which are presented to it, in real time. These measurement selections may also vary in real time, which provides flexibility of the model solution and the model 's uses. PlantStar passes the measurements that are available at the plant, to the dynamic observer model through a "plugin" module, which has been developed to incorporate the observer model and utilise the PlantStar control platform. In an on-line situation, the model will track the plant's behaviour and continuously update its position in real-time to ensure it follows the plant closely. This model would then be able to run simulations of the plant in parallel and could be used as a training facility for new operators, while in a real-time situation it could provide estimates of unmeasurable variables throughout the plant. An example of some of these variables are the flotation rate constants of minerals throughout the plant, which can be estimated in real time by the extended Kalman filter. The model could also be used to predict future plant conditions based on the current plant state , allowing for case scenarios to be performed without affecting the actual plant's performance. Once the dynamic observer model and "plugin" module were completed, case scenario simulations were performed using a measured data set from the plant as a starting point because real-time data were unavailable as the model was developed off-site . / Thesis (M.Sc.Eng.)-University of Natal, Durban, 2004.

For many years the problem of separating aliphatic and aromatic compounds has been at the forefront of the petroleum and oil refining industries. This separation is often effected using liquid-liquid extraction or extractive distillation. Both of these processes require the addition of a solvent to bring about separation. The aims of this work were to investigate the use of "mixed" solvents, such as those used in the Arosolvan process, for their application in liquid-liquid extraction and extractive distillation as well as to provide related thelmodynamic data for systems containing mixed solvents. In the last part of this work, a computer program was developed to theoretically predict the effectiveness of a number of solvents on a user-defined separation. The solvents used for liquid-liquid extraction were chosen based on their similarities to those in the Arosolvan process and were of the form, {N-methyl-2-pyrollidone (NMP) + glycerol, a glycol or water} where the glycol was either monoethylene glycol (MEG), diethylene glycol (DEG) or triethylene glycol (TEG). The additives were combined in various mixing ratios to NMP to determine a mixing ratio for which the effect of the solvent is possibly optimized (a list of all solvents and mixing ratios used are presented in this work). Solvent selectivity and the range of compositions over which separation could occur determined the effectiveness of the solvents. This work dealt with the separation of n-hexane and toluene. In order to determine the selectivity and range of compositions, the liquid-liquid equilibria (LLE) of systems containing n-hexane + toluene + solvent had to be determined. LLE was measured using a simple equilibrium cell at 298 K and 1 atm. The phase separation boundaries (binodal curves) were determined using a titration method. The results obtained in this work showed an increase in the range of compositions over which the mixture of n-hexane and toluene could be separated (i.e a larger range of mixing ratios over which these components could be separated from each other) from the pure NMP solvent to the mixed solvent cases. This implies that there is a The range of compositions over which separation could be affected is given (for the solvents) in descending order: NMP + 50% glycerol> NMP + 10% water > NMP + 30% MEG > NMP + 5% water > NMP + 30% glycerol> NMP + 10% glycerol > NMP + 10% MEG > NMP + 10% DEG > NMP + 10% TEG > NMP + 5% DEG > 100% NMP. The selectivities of the solvents showed a remarkable increase from the pure NMP case to the mixed solvent cases. The maximum selectivity obtained for the NMP + 10% DEG system was over 1200 compared to a maximum selectivity of just 6 for the pure NMP system. The maximum selectivities obtained in descending order were as follows: NMP + 10% DEG > NMP + 10% TEG > NMP + 10% glycerol > NMP + 10% MEG > NMP + 30% MEG > NMP + 50% glycerol > NMP + 10% water > NMP + 5% water > NMP + 30% glycerol > NMP + 5% DEG > 100% NMP. The binodal curves were modelled using the Hlavaty, ,8-density and log-y functions. The maximum standard deviations obtained were 0.075, 0.078 and 0.05 for each of the functions respectively. The equilibrium data was modelled using the UNIQUAC and NRTL thermodynamic models and showed excellent agreement. This work showed better agreement to the NRTL functions due to the fact that the non-randomness parameter, a ij , may be chosen arbitrarily. The results obtained in this work indicate that the use of mixed solvents greatly increases the effectiveness ofNMP used for the separation of n-hexane and toluene. It is suggested that further studies be performed on a wider range of aliphatic and aromatic compounds in order to determine whether this is a generic behaviour or just true for n-hexane and toluene. The effectiveness of each solvent for extractive distillation was determined by its separation factor. In order to determine separation factors, the activity coefficients at infinite dilution (IDACs) had to be measured. This was done using a gas-liquid chromatography technique. The solvents employed in this study were NMP, Glycerol, MEG, TEG, NMP + 10% glycerol, NMP + 10% MEG, NMP + 10% DEG, NMP + 10% TEG. The solutes used were: pentane, heptane, hexane, toluene and benzene. The separation factors were determined for each alkane/aromatic pair per solvent. The pure solvent cases were then compared to the mixed solvent cases. The mixed solvents did not show results as promising for extractive distillation applications as they did for liquid-liquid extraction. TEG displayed the best selectivities for each of the alkane/aromatic separations except for the heptane/benzene pair, for which NMP + 10% glycerol proved to be the most effective solvent. When compared to the results obtained from the original UNIF AC model, the IDACs obtained in this work showed up to a 99% deviation. This is due to the fact that the model does not work well for all types of molecules and does not predict the equilibrium of "unlike" molecules adequately. It is suggested that other mixing ratios and different solvents be used to further investigate the effectiveness of mixed solvents for extractive distillation applications. It is further recommended that a computer aided data logging system be developed to determine residence times. This would not only provide more accurate results, but also provide a database for future reference. The computer program that was developed using the original UNIF AC method contains a database of 28 commonly used industrial solvents. This program enables the user to compare graphically the effectiveness of each of the solvents on the desired separation. Due to the limitations of the original UNIF AC method, the program does not work well for all types of molecules. However, the model can be changed without altering the prografnming structure to include a modified version of the UNIFAC model depending on the users needs. The program although written from an extractive distillation standpoint can be extended to include liquid-liquid equilibrium predictions. The main benefit of such a program is to eliminate time-consuming experimental work required to narrow down a long list of solvents required for a particular separation by theoretically predicting the best solvents for the job. The solvent database can also be expanded when new solvents become available or the user needs change / Thesis (M.Sc. Eng)-University of Natal, Durban, 2003.

This study investigates and compares the environmental burdens of two different methods for producing potable water by using the environmental life cycle assessment (LCA). The first method, for the production of potable water, is used by Umgeni Water at their Wiggins Waterworks and it involves conventional processes. The second method is based on a South African membrane technology and currently it is used in three pilot plants around the country. The life cycle concept gives the means understand the environmental impacts associated with a product. process or activity by considering all life-cycle stages, from cradle- to-grave. Formal methodologies for conducting such studies have been developed and in this project the methodological framework endorsed by the International Organisation for Standardisation (ISO) 14040 series of standards has been used. By using this methodology and by tracing all processes involved in the production of potable water, it was found that the main contribution towards the environmental burdens of potable water is due to electricity generation. This conclusion is valid for both methods in vestigated. and as a result the recommendations focus on increasing the energy efficiency of waterworks in order to increase their overall environmental performance. / Thesis (M.Sc.Eng.)-University of Natal, Durban, 2001.

A waste acid stream is being produced by a local petrochemical company (SASOL) at a rate of 10 000 -12 000 tons per annum and contains approximately 44-mole % butyric acid, 20 % isobutyric acid and 10 % valeric acid. Whilst this stream is currently being incinerated, SASOL has requested an investigation into the possibility of separating and purifying butyric acid and isobutyric acid from this waste acid stream. The goal of this project was to determine a separation and purification route for butyric acid and isobutyric acid from SASOL'S waste acid stream. In order to achieve this, vacuum distillation and freeze crystallization were chosen for the recovery and purification of the acids respectively. Vapour-liquid equilibrium data for key component pairs present in the waste acid stream (propionic acid + butyric acid, isobutyric acid + butyric acid, butyric acid + isovaleric acid and butyric acid + hexanoic acid) were experimentally determined in a dynamic VLE still. The measured VLE data was successfully correlated us ing the gamma-phi approach. with the NRTL activity coefficient model representing the liquid phase and the virial equation of state describing the vapour phase. Using these equations. the VLE data obtained from the experimental work was then regressed to provide interaction coefficients for the NRTL model. which were then used in the Hysys process simulator to explore a range of design alternatives for distillation. Hysys simulations showed that greater than 80 % butyric acid and isobutyric acid can be recovered from the waste acid stream in a single distillation column containing 18 theoretical stages and an optimum reflux ratio of 3.8. The simulation was performed at a pressure of 58kPa and a maximum operating tempe rature of 150°C. Batch distillation experiments performed in a batch rectification column at 250kPa recovered more than 90% of both the butyric acid and isobutyric acid from a 450ml sample of the waste acid stream. A subsequent batch experiment concentrated the recovered acids into a distillate containing more than 95 % butyric acid and isobutyric acid combined. To investigate freeze crystallization as a suitable operation for purifying butyric acid and isobutyric acid a solid-liquid phase equilibrium curve for the system was generated us ing the Van Hoft equation. The generated curve showed that butyric acid and isobutyric acid could be theoretically purified (>98%) by operating two crystallizers at -20°C and -55°C respectively. A simple freeze crystallization experiment produced butyric acid with greater than 94% purity. An economic feasibility study conducted on the process showed that separation and purification of the acids by this process (distillation and crystallization) could create a business opportunity with revenue of approximately R47 million per annum. Preliminary estimates for capital investment amounted to approximately R5.4 million. for which the payback period was estimated at less than one year. / Thesis (M.Sc.Eng.)-University of Natal, Durban, 2001.

At present, South Africa has inadequate technology to destroy its hazardous waste, with approximately 18000 litres of chlorinated hazardous waste stored in this country. Approximately 800 tons of banned or obsolete chemicals are to be sent to Pontypool. Wales, for incineration, at a considerable cost. Because of the toxic nature of chlorinated waste and their long-term effects on the environment , a sustainable method of dealing with this type of waste is essential. Gas phase destruction of methylene chloride, trichlorobenzene and lindane by pyrolysis (i.e. heating in the absence of oxygen) was attempted. Destruction was effected by high temperature thermal degradation of molecules into free radicals. These radicals then combine to form hydrogen chloride and carbon as major products. This method was chosen so as to eliminate the possible formation of highly toxic oxygenated derivatives such as polychlorinated dibenzofurans and dibenzodioxins that can be formed during incineration if strict control is not excercised. The reactor assembly was built in the Department of Chemical Engineering at the University of Natal. 11 incorporates aspects of many different previously designed reactors, as discussed in the text. Heat for the reactions was supplied by induction. A high frequency induction unit supplied current to a copper coil. The resulting magnetic field induced current to flow in a susceptor housed within the copper coil. The susceptor in this case was a graphite tube, which served as both the heating element and the thermal radiation source, in addition to forming the walls of the reaction zone. Up and down stream processes were designed and experiments were carried out in which reaction temperatures (348-1400°C) and residence times (1.3-5.6 seconds) were varied. Destruction efficiencies of 100% and 99.99% were obtained for methylene chloride and trichlorobenzene respectively, with inert argon used as the carrier gas. These destruction efficiencies comply with the 99.99% stipulated by the United States Conservation and Recovery Act. A cause for concern was the formation of chlorinated benzenes and naphthalenes. Destruction of lindane proved unsuccessful due to limitations in the vapourisation and feed system and will have to be investigated further. The method of induction heating was evaluated to be 98.9% thermally efficient. Raw material and utility consumption per ton of waste destroyed by the pyrolysis process was compared to values for incineration as well as the plasma arc and catalytic extraction processes. Consumption for pyrolysis compares favourably with all three processes and suggests that the process could be competitive. Claims to the success of the technology on a wide scale are limited by the small number of compounds that were successfully pyrolysed. Results do however indicate much promise for this technology to be used as a fi nal chlorinated waste destruction unit on an existing process. Modifications to the existing reactor to improve product recovery and analys is will allow for temperature and residence time optimisation for a variety of wastes. Additional in strumentation and process control will allow for kinetic studies to be undertaken in future. This project should be considered as the first step in an ongoing series of research and subsequent improvements to the technology presented here. / Thesis (M.Sc.Eng.)-University of Natal, Durban, 2001.