Biological sulphate reduction (BSR) has been identified as a promising alternative technology for the treatment of acid mine drainage. BSR is a process that uses sulphate reducing bacteria to reduce sulphate to sulphide using substrates as nutrients under anaerobic conditions. The performance of BSR is dependent on several factors including substrate, pH, temperature and hydraulic retention time (HRT). In a quest to find a cost effective technology, Mintek conducted bench-scale tests on BSR that led to the commissioning of a pilot plant at a coal mine in Mpumalanga province, South Africa. This current study forms part of the ongoing tests that are conducted to improve Mintek’s process. The purpose of this study was to investigate the robustness of Mintek’s process and to develop a tool that can be used to predict the process’ performance with varying pH, temperature and HRT. Design Expert version 11.1.2.0 was used to design the experiments using the Box-Behnken design. In the design, pH ranged from 4 to 6, temperature from 10 °C to 30 °C and HRT from 2 d to 7 d with sulphate reduction efficiency, sulphate reduction rate and sulphide production as response variables. Experiments were carried out in water jacketed packed bed reactors that were operated in a down-flow mode. The reactors were packed with woodchips, wood shaving, hay, lucerne straw and cow manure as support for sulphate reducing bacteria (SRB) biofilm. Cow manure and lucerne pellets were used as the main substrates and they were replenished once a week. These reactors mimicked the pilot plant. The data obtained were statistically analysed using response surface methodology. The results showed that pH did not have a significant impact on the responses (p>0.05). Temperature and HRT, on the other hand, greatly impacted the process (p<0.05) and the interaction between these two factors was found to be strong. Sulphate reduction efficiency and sulphate reduction rate decreased by over 60 % with a decrease in temperature 30 °C to 10 °C. Generally, a decrease in sulphide production was observed with a decrease in temperature. Overall, a decrease in HRT resulted in a decline of sulphate reduction efficiency and sulphide production but favoured sulphate reduction rate. This study demonstrated that Mintek’s process can be operated at pH as low as 4 without any significant impact on the performance. This decreases the lime requirements and sludge production during the pre-neutralisation stage by close to 50 %. There was, however, a strong interaction between temperature and HRT which can be used to improve the performance especially during the winter season. / Dissertation (MEng)--University of Pretoria, 2020. / Chemical Engineering / MEng / Unrestricted

Hydrocolloid stabilizers are commonly used in ice cream formulations to provide body and reduce ice crystal growth during storage. We conducted a retail survey of 65 different vanilla ice cream brands and found the majority of manufacturers primarily use 1 or more of 4 different hydrocolloid sources: guar gum, carrageenan, locust and carob bean gum, or cellulose gum or gel. However, many consumers view hydrocolloids as unnatural, and the presence of hydrocolloids on an ingredient declaration may negatively affect purchase intent. Our survey of 705 consumers showed significant differences in purchase intent for vanilla ice cream, based on ingredient declarations containing different hydrocolloid stabilizers. A response surface central composite design was used to optimize ice cream stability using combinations of tapioca flour, carob bean gum, and citrus fiber, 3 consumer-preferred hydrocolloid stabilizers. Instrumental evaluations considered the dependent variables mix viscosity, ice cream hardness and toughness, melt-rate, and ice crystal size. A trained sensory panel also rated iciness, melt-rate, ease of breakdown in the mouth, and vanilla intensity. Each of the dependent variables from the trained panel and instrumental analysis were measured before and after a 3-week accelerated temperature cycling test. A regression analysis of the central composite design data combined instrumental and trained-panel results to compute a response surface based on the regression equation of each attribute. Using the response surface, 3 different optimized mix formulations were determined. The 3 different mixes were optimized using: 1) all dependent variables evaluated, 2) only sensory iciness scores, and 3) ice crystal size only. An untrained consumer panel evaluated samples before and after temperature cycling test, and rated vanilla ice creams prepared from all 3 optimized mixes against a control ice cream, prepared with a natural commercial stabilizer blend. The uncycled products prepared using optimized stabilizer blend were at statistical parity with the control product for overall acceptance, purchase likelihood, preference, sweetness and vanilla intensity, rate of melting in the mouth, texture and hardness. One or more of the optimized products were rated significantly better than the control for creaminess and texture. For products subjected to temperature cycling, 1 or more of the optimized products were rated significantly better than control for all attributes except sweetness and vanilla flavor intensity. This research indicates that more consumer-preferred options for ice cream stabilization are available to the ice cream industry, with performance and sensory results equal to other commercial hydrocolloid stabilizer blends.

No description available.

Ferritic stainless steel is typically used in the automotive industry to fabricate welded tube that is plastically deformed for flanging, bending and necking. The effect of welding parameters during autogenous gastungsten arc welding (GTAW) of thin sheet on the weld metal structure and tensile properties were determined. Two grades of ferritic stainless steels, a titanium-containing Grade 441 and a titanium-free molybdenum-containing Grade 436, were used as base metal. Statistical analysis was used to determine the influence of welding parameters on the microstructure of autogenous GTAW welds. The results of Grade 441 indicated that the welding speed and peak welding current had a statistically significant influence on the amount of equiaxed grains that formed. For Grade 436, the same welding parameters (welding speed and peak welding current) had a statistically significant influence on the grain size of the weld metal grains. The ductility of a tensile test coupon machined parallel to the weld direction, for both base metal grades, was unaffected by the welding parameters or the weld metal microstructure. The elongation was determined by the amount of weld metal in the gauge area of a tensile coupon. The titanium content of the base material seems to have the most significant effect on the formation of equiaxed grains. / Dissertation (MEng)--University of Pretoria, 2019. / Materials Science and Metallurgical Engineering / MEng / Unrestricted

The purpose of this study is to develop a series of response surfaces that define critical outcomes for welding in Al 7075 based on the tool geometry of a convex scrolled shoulder step spiral (CS4) friction stir processing tool. These response surfaces will be used to find critical minimums in forces which will decrease the required power input for the process. A comprehensive parameterization of the tool geometry is defined in this paper. A pilot study was performed to determine the feasibility of varying certain geometric features. Then a screening experiment eliminated those geometric features that were not as significant in determining the response surfaces. A central composite design with the five most important geometric features was used in order to develop response surfaces for nine different response variables. The nine response variables are the longitudinal, lateral and axial forces; the tool temperature, the spindle torque, the amount of flash, the presence of defects, the surface roughness and the ledge size. By using standard regression techniques, response surface equations were developed that will allow the user to optimize tool geometries based on the desired response variables. The five geometric features, the process parameters and several of their interactions were found to be highly significant in the response surfaces.

In today's competitive manufacturing industries, companies strive to reduce manufacturing development costs and lead times in hopes of reducing costs and capturing more market share from early release of their new or redesigned products. Tooling lead time constraints are some of the more significant challenges facing product development of advanced free-form components. This is especially true for complex designs in which large dies, molds or other large forming tools are required. The lead time for tooling, in general, consists of three main components; material acquisition, tool design and engineering, and tool manufacturing. Lead times for material acquisition and tool manufacture are normally a function of vendor/outsourcing constraints, manufacturing techniques and complexity of tooling being produced. The tool design and engineering component is a function of available manpower, engineering expertise, type of design problem (initial design or redesign of tooling), and complexity of the design problem. To reduce the tool design/engineering lead time, many engineering groups have implemented Computer-Aided Design, Engineering, and Manufacturing (CAD/CAE/CAM or CAx) tools as their standard practice for the design and analysis of their products. Although the predictive capabilities are efficient, using CAx tools to expedite advanced die design is time consuming due to the free-form nature and complexity of the desired part geometry. Design iterations can consume large quantities of time and money, thus driving profit margins down or even being infeasible from a cost and schedule standpoint. Any savings based on a reduction in time are desired so long as quality is not sacrificed. This thesis presents an automated tool design methodology that integrates state-of-the-art numerical surface fitting methods with commercially available CAD/CAE/CAM technologies and optimization software. The intent is to virtually create tooling wherein work-piece geometries have been optimized producing products that capture accurate design intent. Results show a significant reduction in design/engineering tool development time. This is due to the integration and automation of associative tooling surfaces automatically derived from the known final design intent geometry. Because this approach extends commercially available CAx tools, this thesis can be used as a blueprint for any automotive or aerospace tooling need to eliminate significant time and costs from the manufacture of complex free-form components.

Within statistical methods, design of experiments (DOE) is well suited to make good inference from a minimal amount of data. Two types of designs within DOE are screening designs and optimization designs. Traditionally, these approaches have been necessarily separated by a gap between the objectives of each design and the methods available. Despite being so separated, in practice these designs are frequently connected by sequential experimentation. In fact, from the genesis of a project, the experimentor often knows that both designs will be necessary to accomplish his objectives. Due to advances in the understanding of experimental designs with complex aliasing and their analysis, a current topic within statistics is how to desegregate these methods into a more unified and economical approach. This project is one treatment of that issue.

Errors of computational simulations in design of a high-speed civil transport (HSCT) are investigated. First, discretization error from a supersonic panel code, WINGDES, is considered. Second, convergence error from a structural optimization procedure using GENESIS is considered along with the Rosenbrock test problem. A grid converge study is performed to estimate the order of the discretization error in the lift coefficient (CL) of the HSCT calculated from WINGDES. A response surface (RS) model using several mesh sizes is applied to reduce the noise magnification problem associated with the Richardson extrapolation. The RS model is shown to be more efficient than Richardson extrapolation via careful use of design of experiments. A programming error caused inaccurate optimization results for the Rosenbrock test function, while inadequate convergence criteria of the structural optimization produced error in wing structural weight of the HSCT. The Weibull distribution is successfully fit to the optimization errors of both problems. The probabilistic model enables us to estimate average errors without performing very accurate optimization runs that can be expensive, by using differences between two sets of results with different optimization control parameters such as initial design points or convergence criteria. Optimization results with large errors, outliers, produced inaccurate RS approximations. A robust regression technique, M-estimation implemented by iteratively reweighted least squares (IRLS), is used to identify the outliers, which are then repaired by higher fidelity optimizations. The IRLS procedure is applied to the results of the Rosenbrock test problem, and wing structural weight from the structural optimization of the HSCT. A nonsymmetric IRLS (NIRLS), utilizing one-sidedness of optimization errors, is more effective than IRLS in identifying outliers. Detection and repair of the outliers improve accuracy of the RS approximations. Finally, configuration optimizations of the HSCT are performed using the improved wing bending material weight RS models. / Ph. D.

Traditional single stage design optimality procedures can be used to efficiently generate data for an assumed model y = f(x^(m),b) + &#949;. The model assumptions include the form of f, the set of regressors, x^(m) , and the distribution of &#949;. The nature of the response, y, often provides information about the model form (f) and the error distribution. It is more difficult to know, apriori, the specific set of regressors which will best explain the relationship between the response and a set of design (control) variables x. Misspecification of x^(m) will result in a design which is efficient, but for the wrong model. A Bayesian two stage design approach makes it possible to efficiently design experiments when initial knowledge of x^(m) is poor. This is accomplished by using a Bayesian optimality criterion in the first stage which is robust to model uncertainty. Bayesian analysis of first stage data reduces uncertainty associated with x^(m), enabling the remaining design points (second stage design) to be chosen with greater efficiency. The second stage design is then generated from an optimality procedure which incorporates the improved model knowledge. Using this approach, numerous two stage design procedures have been developed for the normal linear model. Extending this concept, a Bayesian design augmentation procedure has been developed for the purpose of efficiently obtaining data for variance modeling, when initial knowledge of the variance model is poor. / Ph. D.

Impaired reproduction studies with Poisson responses are among a growing class of toxicity studies in the biological and medical realm. In recent years, little effort has been focused on the development of efficient experimental designs for impaired reproduction studies. This research concentrates on two areas: 1) the use of Bayesian techniques to make single regressor designs robust to parameter misspecification and 2) the extension of design optimality methods to the k-regressor model. The standard Poisson model with log link is used. Bayesian designs with priors on the parameters are explored using both the D and F-optimality criteria for the single regressor Poisson exponential model. Since these designs are found via numeric optimization techniques, Bayesian equivalence theory functions are derived to verify the optimality of these designs. Efficient Bayesian designs which provide for lack-of-fit testing are discussed. Characterizations of D, D_s, and interaction optimal designs which are factorial in nature are demonstrated for models involving interaction through k factors. The optimality of these designs is verified using equivalence theory. In addition, augmentations of these designs that result in desirable lack of fit properties are discussed. Also, a structure for fractional factorials is given in which specific points are added one at a time to the main effect design in order to gain estimability of the desired interactions. Robustness properties are addressed as well. Finally, this entire line of research is extended to industrial exponential models where different regressors work to increase and/or decrease a count data response produced by a process. / Ph. D.