Optimisation attempts within the petrochemical industry have led to interest in alternate solvents. The most widely used commercial solvents for the separation of hydrocarbons, by extractive distillation, are N-methylpyrrolidone and sulfolane. There has also been reference made to other solvents, such as N-formylmorpholine and the ethylene glycols [mono-, di-, tri- and tetra], being used. The alternate solvents proposed for this study were N-formylmorpholine, triethylene glycol and diethylene glycol. Infinite dilution activity coefficients, γ∞, provided a means of comparing the ease of separation of the different solutes using different solvents in extractive distillation. There is a substantial database of γ∞ measurements for systems involving N-methylpyrrolidone and hydrocarbons. A fairly large data set of γ∞ values of hydrocarbons in N-formylmorpholine has also been measured. Very little work has been conducted on the γ∞ values of hydrocarbons in either diethylene glycol or triethylene glycol. Gas liquid chromatography is one of the more common methods used to measure γ∞. To enable the measurement of γ∞ at higher temperatures, a pre-saturator was installed prior to the column. This ensured that the carrier gas entering the column was saturated with solvent and prevented the elution of solvent from the column. The γ∞ values of 25 solutes; including n-alkanes, alk-1-enes, alk-1-ynes, alcohols and aromatics; were measured at temperatures of 333.15, 348.15 and 363.15 K. The γ∞ measurements in N-formylmorpholine were used to verify this experimental set up and technique. Once the experimental set up had been proven, γ∞ in N-methylpyrrolidone, triethylene glycol and diethylene glycol were measured. Selectivities and capacities, at infinite dilution, of several solute combinations in the four solvents were then compared. In a few of these separation cases, the alternative solvents appeared to have better separation performance than N-methylpyrrolidone. The γ∞ values of three of the solutes in N-formylmorpholine and N-methylpyrrolidone were also measured using the novel cell design and measurement procedure suggested by Richon. It was found that this new technique required further development for the case of volatile solvents, since the results obtained using this technique did not compare favourably with the literature data. / Thesis (M.Sc.)-University of KwaZulu-Natal, Durban, 2012.

No abstract available. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2009.

Olefin (alkene) metathesis can be used for the redistribution of carbon bonds to produce preferred higher range carbon number olefins (C10-C16) from low value medium chain olefins such as 1-hexene. In this study, the performance of various heterogeneous, tungsten based catalysts was investigated for the gas-phase linear cross metathesis of 1-hexene. The tested catalysts included tungsten trioxide on alumina, at various specific loadings, 8wt% tungsten trioxide on silica as well as 8wt% tungsten trioxide on silica, at various specific loadings of a potassium promoter. An existing lab-scale, fixed bed metathesis reactor system was used to conduct the required experimental work under steady state conditions. The experimental plan used for conducting the investigation of the alumina-supported catalyst was a combination of the One-Variable-At-a-Time (OVAT) approach as well as the factorial design method. The domain of the range under investigation was 8-20% for the specific loading of tungsten trioxide, 420-500oC for reaction temperature and 30-80 mol% for the feed composition of 1-hexene with constant space time. A half factorial experimental design was used for the WO3/SiO2 and potassium doped WO3/SiO2 catalysts. The potassium loadings of the 8wt% WO3/SiO2 catalyst were between 0.05-0.5wt%. Reaction temperatures of between 420-460°C were used together with 1-hexene feed compositions of between 60-80mol% and space times of 200-400g.min.mol-1. The WO3/Al2O3 catalyst was found to be an inferior catalyst for the metathesis of 1-hexene at all combinations of specific loadings and operating conditions tested. The highest yields obtained for both the detergent range olefins (C10-C16) and primary metathesis product decene (C10) were less than 1.5%. Changes made to the calcination temperature, calcination time and pH of the impregnation solution during the catalyst preparation stages had no significant effect on the yields produced. The WO3/Al2O3 catalyst was found to behave more as an isomerization catalyst rather than one for the metathesis reaction. The optimum reaction conditions determined when investigating the 8wt% WO3/SiO2 catalyst were found to be a reaction temperature of 460°C, a feed gas composition of 60mol% 1-hexene and a space time of 400g.min.mol-1. The conversion of 1-hexene, the yield of the detergent range olefins and the yield of decene at the above mentioned reaction conditions were 82%, 8.30% and 5.92% respectively Doping of the WO3/SiO2 catalyst with potassium was found to be successful in reducing the amount of isomerization and increasing the yields of both the detergent range olefins and decene by approximately 1.5% and 2% respectively when doping with 0.1 and 0.5wt% potassium. The experimental runs conducted were at the optimum reaction conditions obtained using the 8wt% WO3/SiO2 catalyst. The highest selectivity of the detergent range olefins (24.64%) and decene (23.62%) was obtained when using the 0.5wt% K doped WO3/SiO2 catalyst. At the optimised reaction conditions, the 0.5wt% potassium loading on WO3/SiO2 performed the best. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2013.

The gas-liquid reaction between chlorodifluoromethane (R-22) and methanol, in the presence of sodium hydroxide, was investigated in an isothermal, stirred, semi-batch reactor. The objective of the study was to develop a model for the reaction and to identify the kinetic parameters. Reactor temperature was varied from 283 to 303 K, with inlet R-22 partial pressures between 40.5 and 60.8 kPa (absolute). Solutions containing sodium hydroxide concentrations of between 1.5 and 2.5 mol·dm-3 were charged into the reactor prior to each experiment. Preliminary investigations using the R-22-methanol system revealed that stainless steel was an inappropriate choice of material for the reactor as it displayed catalytic tendencies toward trimethyl orthoformate formation. Consequently, the reactor was constructed from glass and was equipped with an internal cooling coil, a single heating jacket and a temperature control unit. Liquid samples that were withdrawn from the reactor were degassed under vacuum to remove residual chlorodifluoromethane, and thereby inhibit further reaction. Spectrophotometry was used to analyze the liquid samples to determine the concentration of chloride ions in solution. The products obtained were difluorodimethyl ether (major product) and trimethyl orthoformate (by-product) as well as sodium chloride and sodium fluoride salts. Difluorodimethyl ether is a potential replacement for ozone depleting CFC refrigerants. A Box-Behnken experimental design was used to investigate the effect of reaction conditions on the product distribution. Variations in the reaction temperature, initial concentration of sodium hydroxide and inlet partial pressure of R-22 were considered. The modeling of the gas-liquid reactions was based on the -dehydrohalogenation mechanism. Since gas solubility in a liquid decreases in the presence of dissolved salts, the "salting-out" effect on mass transfer was included in the reactor model. Sechenov coefficients for sodium chloride and sodium fluoride were combined to give a salt Sechenov coefficient Ksalt . It was known from the literature that the presence of precipitated salts causes inefficient mixing and inhibits mass transfer, particularly in this system due to the relatively low salt solubilities in methanol. This mixing effect was also included in the appropriate mass transfer terms of the reactor model. The experimental data was fitted to a proposed kinetic scheme. Kinetic parameters for each of the proposed reactions, the Sechenov ‘salting out’ coefficients and the mixing parameter were obtained through the use of a non-linear, least-squares optimization algorithm. For the kinetic study, activation energies of 89.12 and 45.83 kJ·mol-1 were obtained for the difluorodimethyl ether and trimethyl orthoformate formation reactions, respectively, with a Sechenov salt coefficient of 0.712 and a mixing parameter of 22.43. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2013.

One of the consequences of the social and economIc change due to industrialisation is the generation of industrial wastewater which requires treatment before being released into the natural aquatic environment. The municipality has wastewater treatment plants which were initially designed for the treatment of domestic wastewater. The presence of industrial wastewater in these treatment plants introduces various difficulties in the treatment process due to the complex and varying nature of the industrial wastewater. A means needs to be developed, that will allow the municipality to evaluate if a wastewater treatment plant can adequately treat a particular composition or type of wastewater to a quality suitable for release to the environment. Developing a simulation model for a wastewater treatment plant and calibrating it against plant operating data will allow the response of the wastewater treatment plant to a particular wastewater to be evaluated. In this study a model for the Mariamidge Wastewater Treatment Plant is developed in the WEST (Worldwide Engine for Simulation, Training and Automation) software package. The sources of data for modelling were laboratory experiments, historical data from the municipal laboratory and modelling of experiments. Dynamic input files representing the properties of the influent wastewater were generated by characterising the influent wastewater through the use of batch respirometric tests and flocculation filtration on composite samples of wastewater. Kinetic and stoichiometric coefficients of the model were determined from batch respirometric tests on wastewater and activated sludge, and simulation of the batch respirometric experiment. To make the model plant-specific it is calibrated against plant operating data. Influent characterisation and reliable ASM3 model parameters were determined from the respirometric batch test and modelling of experiments. The resulting plant model was able to closely predict the trends of the effluent COD concentration in the plant. Hence it was concluded that the use of laboratory experiments, historical data from the municipal laboratory and modelling of experiments in order to generate information for the modelling of wastewater treatment plants makes up a methodology which can be adopted and improved by additional experiments. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2008.

The objective of this research project is to develop new software tools capable of operational optimisation of existing, large-scale water distribution networks. Since pumping operations represent the main operating cost of any water supply scheme, the optimisation problem is equivalent to providing a new sequence for pumping operations that makes better use of the different electricity tariff structures available to the operators of distribution systems. The minimisation of pumping costs can be achieved by using an optimal schedule that will allow best use of gravitational flows, and restriction of pumping to low-cost power periods as far as possible. A secondary objective of the operational optimisation is to maintain the desired level of disinfectant chlorine at the point of delivery to consumers. There is a steady loss of chlorine with residence time in the system. If the level drops too low there is a risk of bacterial activity. Re-dosage points are sometimes provided in the network. Conversely, too high a level produces an unacceptable odour. The combinatation of dynamic elements (reservoir volumes and chlorine concentration responses) and discrete elements (pump stati and valve positions) makes this a challenging Model Predictive Control (MPC) and constrained optimisation problem, which was solved using MINLP (Mixed Integer Non-linear Programming). The MINLP algorithm was selected for its ability to handle a large number of integer choices (valves open or shut / pumps on or off in this particular case). A model is defined on the basis of a standard element, viz. a vessel containing a variable volume, capable of receiving multiple inputs and delivering just two outputs. The physical properties of an element can be defined in such a way as to allow representation of any item in the actual network: pipes (including junctions and splits), reservoirs, and of course, valves or pumps. The overall network is defined by the inter-linking of a number of standard elements. Once the network has been created within the model, the model predictive control algorithm minimises a penalty function on each time-step, over a defined time horizon from the present, with all variables also obeying defined constraints in this horizon. This constrained non-linear optimization requires an estimate of expected consumer demand profile, which is obtained from historical data stored by the SCADA system monitoring the network. Electricity cost patterns, valve positions, pump characteristics, and reservoir properties (volumes, emergency levels, setpoints) are some of the parameters required for the operational optimisation of the system. / Thesis (M.Sc.Eng.)-University of Natal, Durban, 2004.

No abstract available. / Thesis (M.Sc.Eng.)-University of Durban-Westville, 2000.

The primary focus of this study is the measurement and modeling of binary and ternary VLE data. The measurements of binary and ternary systems were undertaken on a fully automated dynamic VLE apparatus. The glass dynamic VLE still was modified to handle pressures ranging from 0 to 500 kPa, however, the safest maximum pressure to which tests had been conducted was 350 kPa. Thus, this limit was not to be exceeded during the measurement of experimental data. The systems under investigation included the binary and ternary combinations of the following chemicals: n-hexane, 1-hexene and n-methyl-2-pyrrolidone (NMP) at isothermal conditions. A test system consisting of ethanol + cyclohexane was measured at 40 kPa, as well as the system of 1-hexene + NMP at 363.15 K and n-hexane + NMP at 363.15 K. Published literature data for these test systems were employed to verify the measured data for the test systems complied with thermodynamic consistency. All other data constitutes new data, currently unavailable in literature. The following isotherms were measured: 1) 1-hexene (1) + NMP (2) at 323.15, 343.15, 353.15 and 363.15 K 2) n-hexane (1) + NMP (2) at 353.15, 363.15, 378.15 and 383.15 K 3) 1-hexene (1) + n-hexane (2) at 343.15, 363.15 and 373.15 K, and 4) 1-hexene (1) +n-hexane (2) + NMP (2) at 363.15 K All system measurements were carried out on the glass low-to-medium pressure VLE still of Lilwanth (2011), with the exception of the test system ethanol + cyclohexane, which was carried out on the low pressure VLE glass still of Hirawan (2007). The two VLE stills, utilized to carry out measurements in this work, can operate isobarically and isothermally. The temperature on the stills of Hirawan (2007) and Lilwanth (2011) were controlled to within ±0.425 and ±0.089 K respectively and the accuracy of pressure control is to within ±0.320 and ±0.440 kPa respectively. In addition, for the calibration of the various systems: ethanol + cyclohexane, 1-hexene + NMP, n-hexane + NMP, 1-hexene + n-hexane and 1-hexene + n-hexane + NMP, the accuracies are: ±0.002, ±0.0034, ±0.0033, ±0.0066 and ±0.0083 of a mole fraction respectively. The binary interaction parameters obtained from modeling the three binary systems were used to predict the ternary system data. Thereafter, the experimentally measured data for the ternary system was then compared to the model prediction, which was completed on Dortmund Data Bank (DDB, 2011). The measured binary data was regressed utilizing the combined and the direct methods. For the direct method, the cubic equations of state (CEoS) used to describe the vapour phase included the Peng-Robinson (1976) and Soave-Redlich-Kwong (1972) equations combined with the mixing rule of Wong and Sandler (1992) in conjunction with the Gibbs excess energy models, namely the NRTL (1968) and UNIQUAC (1975) models, to describe the liquid phase non-idealities. For the combined method, the Gibbs excess energy activity coefficient models mentioned above were employed to represent the liquid phase imperfections and the vapour phase nonidealities were represented by cubic equations of state, as mentioned above, as well as the Hayden and O‟Connell (1975) virial equation of state for the calculation of the virial coefficients. To verify whether the measured data is thermodynamically consistent the point and direct tests were applied. Even though the direct test is a more stringent approach to testing thermodynamic consistency, for the systems 1-hexene + NMP and n-hexane + NMP, the point test was utilized as the primary means by which to quantify the data, as the associative effects of the NMP molecule effect the results obtained. For the system 1-hexene + n-hexane the direct test was used as the primary means to test the consistency of data, as no cross- or self-association is present. After extensive modeling was carried out, it was found that for the systems 1-hexene + NMP and n-hexane + NMP the model which enabled the best fit of the experimental data are the NRTL activity coefficient model in conjunction with the Hayden and O‟Connell virial equation of state (EoS). For the system 1-hexene + n-hexane the overall best fit model is the Peng-Robinson EoS in conjunction with the Wong-Sandler mixing rule and the NRTL activity coefficient model. A single set of binary interaction parameters for each of the three binary systems was obtained (via regression on Aspen Plus®) using the NRTL-HOC models. However, since Aspen Plus® cannot predict ternary system behaviour using the binary interaction parameters of the constituent systems, DDB was utilized. Further, DDB did not have available the HOC virial EoS (for enabling predictions), thus, it was decided to use the ideal gas model for representation of the vapour phase in conjunction with the NRTL activity coefficient model. The use of the ideal gas model does not compromise the integrity of the prediction in any way since the ternary system measurements were carried out in the dilute NMP region. Thus, since the main components in the ternary mixture at any one instant were 1-hexene and nhexane, and these components behave ideally, the ideal gas model is applicable. After the predicted behaviour for the ternary system was compared to the experimental data for the same system, the maximum percentage error encountered between the two data sets is 5%. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2012

In the industrial world where many different types of separation processes are available, liquid-liquid extraction is gaining increasing attention, since it offers many advantages over distillation especially for heat sensitive and azeotropic compounds. Liquid-liquid extraction is an essential separation method that applies to the chemical, petroleum, metallurgy, biotechnology, nuclear and waste management related industries. This separation technique also offers potential means of saving energy, thus making extraction a more economical separation process. The effectiveness of a vibrating plate extractor was previously investigated however limited research was conducted on the effect of tray spacing, solvent-to-feed ratio and agitation level on a vibrating plate extraction column. These parameters affect the hydrodynamics and mass transfer in a vibrating plate extraction column therefore the determination of the optimum process parameters is important in achieving the highest efficiency for the column and this is sought after for a vibrating plate extractor. A toluene-acetone-water system was selected for experimental work to be conducted on the vibrating plate extraction column. This test system for liquid-liquid extraction is a standard system proposed by the European Federation of Chemical Engineering. The research aimed at testing the effect of tray spacing, agitation level (product of amplitude and frequency of vibration), and the ratio of flow rates of the phases on the number of stages in order to optimise the efficiency of a vibrating plate extraction column. For the hydrodynamic experiments the dispersed phase holdup, drop size distribution and Sauter mean diameter were evaluated for varying parameters. For the mass transfer experiments the percentage acetone extracted, number of equilibrium stages, mass transfer coefficient and overall efficiency were determined as well as the dispersed phase holdup, drop size distribution and the Sauter mean diameter for varying parameters. A decrease in the holdup occurred for an increase in the solvent-to-feed ratio and an increase in the agitation level resulted in a decrease in the Sauter mean diameter. Results indicated that lower tray spacing resulted in a higher extraction of acetone. Backmixing in the dispersed phase resulted in higher number of stages. / Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2012.

The South African coal industry produces a large quantity of coal per annum. The rejects from various unit operations, such as spirals, consist of fine coal that joins the plants tailings dam waste. As existing high quality resources become depleted, the need to improve recovery of this fine coal grows. This project investigates the use of a teetered bed separator (TBS); a hindered settling gravity concentration device for fine coal recovery. This device has proven successful in the United Kingdom and in Australian collieries for fine coal separation in the size range between 2mm and 0.3mm. It has also been used for decades as a classifying device for silica sand and tin. The TBS operates in the size range of water-only cyclones and spiral concentrators, and could potentially be used to separate a broader size range of coal fines so as to offer a lower footprint device for the fines recovery section of a plant. Spiral concentrators cannot always be operated efficiently at a separating specific gravity of lower than 1.6; a TBS may also extend the density range for separation and thus improve recovery. The objective of this project was to gain a full understanding of the TBS from fundamental particle interaction and develop a lab scale unit, which is capable of separation to about 0.1mm at optimum conditions. This involved the development of design parameters based on the various distributor plates and flow pattern modelling. The hydrodynamics of the separator were investigated using the Eulerian-Eulerian modelling approach of commercial CFD package, Fluent 6.1. Seven distributor plates of varying aperture size and geometric arrangement were considered. Coal and shale particles, sized between 2mm and 0.038mm with a specific gravity (SG) range of 1.2 to 2.0, were separated using the laboratory scale unit. The results of both the simulations and the laboratory tests were then compared. The simulations revealed that Plate 3 was the best option for implementation. It had an even upward velocity profile compared to the other plates, with minimum wall effects and disturbances. The upward water flow rate (teeter water) was varied experimentally and the composition of the teeter bed, underflow and overflow were analysed using 1.5, 2 and Smm cubic density tracers with an SG range of 1.2-2.0. Analysis of the partition curves of the distributor plates revealed that Plate 3 had the lowest Ecart Probable (Ep) and cut- point densities. The comparison of simulated results and experimental results show that the simulator could predict the distributor plate design with the lowest Ep in practical tests. The simulator could be beneficial when optimising an industrial scale unit, by allowing prediction of improved segregation patterns and thus separation efficiency. / Thesis (M.Sc.)-University of KwaZulu-Natal, Durban, 2005.