Waste streams have increased due to advancements in technology and the increase in the global population, requiring innovative strategies to recover value from them whilst reducing their negative environmental impact and human health hazards. Thus, the increase in waste has led to research focused on circular economies. E-waste is the fastest growing waste stream in the world containing valuable metals that exceed those rich in ore from mines. In Africa, e-waste metal recycling remains largely informal and small scale resulting in inefficient metal recovery, increased negative environmental impact and human health hazards. E-waste metal recycling using pyrometallurgy is limited to secondary smelter feed and there are limited industrial plants dedicated solely for this purpose. While in hydrometallurgy, research in e-waste metal recycling has been largely focused on metal extraction whilst downstream metal recovery processing studies are limited. The strategy often employed in e-waste metal recovery via hydrometallurgy is base metal (BM) extraction before precious metal (PM) recovery due to the high concentration of these metals in e-waste. This results in the production of base metal-rich-leachate solutions. The heterogeneity of these leachate solutions and the high cost of downstream processing requires a multi-disciplinary approach that considers metal recovery selectivity and associated costs. Natural sorbents, chitin and chitosan found in large quantities in industrial food waste and precipitation with sulphides have received much attention due to their high metal recovery efficiencies, metal selectivity, scalable operation and low costs. Chitin and chitosan are mainly sourced from crustacean shell waste and there are limited techno-economic studies on the extraction and production methodology of these polymers. Chitin and chitosan from Black Soldier Fly (BSF) larvae shells, a waste product from BSF farming, is thought to have high adsorptive properties due to their low crystalline index. However, studies on metal adsorption onto chitin and chitosan sourced from BSF larvae and their potential combined application with sulphide precipitation to recover metals from e-waste leachate solutions remains limited. Therefore, the dissertation aimed to develop a cost-effective method of extracting chitin and chitosan from BSF larvae shell waste and investigated the techno-economic feasibility of the application of these polymers in combination with sulphide precipitation for the recovery of base metals from e-waste leachate solutions. The potential application of chitin/chitosan from BSF larvae in e-waste metal recovery may result in a circular economy where solid waste is utilized to produce BSF larvae. While the BSF larvae shell waste generated from BSF larvae production can be used to remediate electronic waste, recovering value from these waste streams while reducing their environmental impact. The cost-effective method for the extraction of chitin and production of chitosan from BSF larvae was investigated by a study into the effects of demineralisation, deproteination, decolourisation, de-acetylation processes on the chitin and chitosan character, metal adsorption performance and techno-economics. Chitin and chitosan were extracted and produced from BSF larvae (Hermetia illucens) using a combination of the processes stated prior. Adsorption studies with the produced chitin and chitosan were conducted on base metals ferrous, ferric, copper and aluminium ions in single and bimetal solutions. The adsorbed metals were then eluted using 0.1 M H2SO4. Precipitation studies were also conducted with various concentrations of copper in a ferrous, copper and aluminium solution. The techno-economic feasibility of the application of the chitin and chitosan and sulphide precipitation with NaHS in PCB leachate solutions was investigated by the development of a model based on the ascertained individual metal recovery performance in the adsorption and precipitation studies. Extracted chitin from BSF larvae was found to be in the alpha form. 4-hour Deproteination of the BSF larvae after liberation with 4 wt % NaOH and de-acetylation of the deproteinated chitin with 40 wt% NaOH was found to produce chitin and chitosan with the highest metal sorption capacities and lowest cost of production. The maximum adsorption capacity for ferrous, ferric, copper and aluminium ions onto chitin from BSF larvae was 2.29 ± 0.0001 mmol/g, 2.07 ± 0.0001 mmol/g, 1.69 ± 0.0001 mmol/g and 1.82± 0.0001 mmol/g respectively. While for chitosan, the maximum adsorption capacity for ferric, copper and aluminium ions was 0.951 ± 0.0012 mmol/g, 1.16 ± 0.0016 mmol/g and 0.961± 0.0013 mmol/g respectively. The order of metal adsorption selectivity for ferrous, ferric, copper and aluminium on chitin from BSF larvae was determined to be Fe2+>Fe3+>Al3+>Cu 2+. While for chitosan it was determined to be Cu2+>Fe3+>Al3+ and at a low pH (below pH of 2) it was observed to be Cu2+>Al3+>Fe 3+. Ferrous ion oxidation to ferric ions was observed during the adsorption of ferrous ions onto the chitin and chitosan. Adsorption of the metals onto chitin and chitosan were best modelled by the Freundlich isotherm and Pseudo 2nd order kinetic model. The adsorption on both polymers was found to be spontaneous, favourable, chemisorption and predominantly surface complexation. Sulphide precipitation with NaHS was observed to be selective towards copper precipitation however co-precipitation with aluminium occurred. The application of chitin and chitosan on the multi-metal synthetic PCB leachate solution resulted in the production of two refined streams respectively. The application of NaHS precipitation seems to be more feasible on the refined streams produced by the application of chitin. The combined application of NaHS and chitin from BSF larvae on the multi-metal synthetic PCB leachate solution showed economic feasibility. The recovery costs were $ 116 per kg metal recovered and an overall gross profit of $ 933/ kg metal recovered. However further economic studies which include consideration of capital costs need to be conducted to conclusively determine the economic feasibility of this downstream metal recovery process. This study shows the potential of chitin and chitosan extracted from BSF larvae to upgrade PCB metal leachate solutions for further downstream processing

This study tests the hypothesis that the incorporation of manganese into cobaltbased/silica (SiO2) catalyst influences the activity and selectivity of the FischerTropsch synthesis via the enhancement of CO adsorption and dissociation. The base catalyst is prepared by impregnation (IMPR) of the SiO2 support with cobalt(II) nitrate hexahydrate followed by calcination at 230 °C under a continuous flow of air. This IMPR approach resulted in the successful formation of well-defined and monodisperse cobalt nanoclusters over SiO2 support material. The catalyst precursor containing 22 wt.% Co/SiO2 were doped with different manganese concentrations using strong electrostatic adsorption (SEA) and IMPR approaches. The physico-chemical properties of the catalyst precursors were characterised using ICP-OES, XRD, TEM, STEM-EDS, H2-TPR, TGA, H2 chemisorption and CO-TPD. The catalytic performances of the reduced catalysts in the Fischer-Tropsch synthesis were tested in a fixed bed reactor at 220 °C and 20 barg with (H2/CO)inlet = 2. The incorporation of small amounts of manganese into the Co/SiO2 catalyst precursor, via SEA, resulted in enhanced catalytic activity and this was attributed to the higher degree of cobalt reduction obtained with these materials. However, high manganese contents resulted in decreased catalytic performance, which may be associated with the blockage of catalytically active, cobalt surface sites. There was no evidence of enhanced CO adsorption and dissociation over Co/SiO2 catalyst as a result of the inclusion of selectively adsorbed manganese. Furthermore, the addition of selectively adsorbed manganese on Co/SiO2 catalyst did not significantly change the CH4, C2-4 and C5+ selectivities. The olefin content was the lowest with very small amount of manganese when compared to unpromoted cobalt catalyst, however, a further increase in manganese concentration resulted in an increased olefin content. Lastly, changing the doping method from SEA to IMPR (without pH adjustment) did not significantly change the activity and selectivity of cobalt catalysts. Thus, the change in the method of doping did not influence the activity and selectivity of the cobalt-based catalysts.

This investigation focuses on the use of derivatives by third-tier South African companies. The market capitalisation of these companies ranges from R65 million to R600 million.

Organic-inorganic halide perovskites have received significant attention as a promising opportunity for low-cost, high-performance optoelectronic devices, such as solar cells, photodetectors, and light-emitting diodes due to their excellent optoelectronic properties and considerable low-cost production processes. There is a great interest to investigate further the possibility of integrating perovskites with other materials for the development of efficient optoelectronic devices. This thesis explores the feasibility of integration of single crystals perovskite and mechanoluminescent materials for the fabrication of pressure sensors. The combination and integration of the unique characteristics of perovskite crystals and mechanoluminescent materials are scarce in the literature. The integration of a perovskite layer and a mechanoluminescent material, such as ZnS:Cu can be used to fabricate a highly sensitive and fast-response device. This thesis will focus on the fabrication of an integrated sensing device made with two main components, a mechanoluminescent material layer, acting as a light-emitting layer, and a single crystal perovskite layer, functioning as a light-harvesting layer. This research covers all the process of the device fabrication and testing experiments under tensile and compression forces. The tests conducted proved that the light emitted by the mechanoluminescent material can be collected by a single crystal perovskite as an integrated sensing device. It is the initial step for further research on the utilization of perovskite materials for sensing devices. / A Thesis submitted to the Department of Industrial and Manufacturing Engineering in partial fulfillment of the requirements for the degree of Master of Science. / 2019 / September 30, 2019. / Includes bibliographical references. / Okenwa Okoli, Professor Directing Thesis; Tarik Dickens, Committee Member; Zhibin Yu, Committee Member.

The goal of this research is to fabricate, characterize and apply the designed polymer based porous piezoelectric material into practical use. In line with the goal, the research objectives are: Select proper materials that have potential to serve as non-toxic, high sensitivity, low cost, light weight, flexible and heat durable piezoelectric materials. Setup fabrication process of piezoelectric foam and enable the piezoelectric material to have piezoelectricity. Characterize the sensitivity, heat durability, charge storage process, capability and actuation behavior of the material. Identify the most significant parameters that affect the performance of the material and optimize the structure design of the material. Apply the piezoelectric material as piezoelectric pressure sensor, impact sensor and other applications. / A Dissertation submitted to the Department of Industrial and Manufacturing Engineering in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / Summer Semester 2018. / July 9, 2018. / Includes bibliographical references. / Changchun Zeng, Professor Directing Dissertation; William S. Oates, University Representative; Richard Liang, Committee Member; Arda Vanli, Committee Member.

Advances in modeling tumor-immune dynamics and therapies offer deeper understandings of the mechanism of tumor evolution in the interdisciplinary field of mathematics and immune-oncology. The main mathematical models are constructed in terms of ordinary differential equations (ODEs) or partial differential equations (PDEs) and analyzed through tools such as Poincaré map, simulation, or numerical bifurcation analysis to understand the system properties. These models succeed in characterizing essential features of tumor behaviors including periodic bursts and the existence of latency. In relationship to practice, these models are also applied to estimate the feasibility and efficacy of treatments ranging from traditional chemotherapy to immunotherapy (ACI). In recent literature, there have been applications of control methods such as optimal control, hybrid automata, and feedback linearization-based tracking control with almost disturbance decoupling in the studies of tumor-immune systems. This thesis presents an attempt to apply the bifurcation control method with washout filters in tumor treatments. This thesis research investigates the dynamics and controlling of the tumor-immune response of immunotherapies, mainly the Adoptive Cell Immunotherapy (ACI) and Interleukin-2 (IL-2). The first part of the thesis presents the nonlinear dynamics of the classic nonlinear ODE tumor-immune model given by Denise Kirschner and John Carl Panetta in 1998. This model concentrates on the nonlinear phenomena of the tumor-immune system under immunotherapies, primarily the bifurcation phenomenon along with the antigenicity of effector cells. Bifurcation phenomena refer to the qualitative changes in system dynamics due to quasi-static changes in system parameters. Antigenicity refers to a capability to distinguish tumor cells from healthy cells. The Kirschner-Panetta model captures a saddle-node bifurcation and a Hopf bifurcation of the tumor-immune response, which separates the tumor evolution into three stages, the “dangerous equilibrium”, the periodic recurrence, and the “safe equilibrium”. The second part applies and analyzes several control strategies on the immunotherapies based on the KP model in order to eradicate tumors or inhibit tumor growths. The first section studies the combination immunotherapy of ACI and IL-2 as an open-loop control system based on Kirschner’s work, which generates a locally asymptotically stable equilibrium. In the second section, this thesis provides a new idea of treatment in the tumor-immune system, that is a closed-loop control strategy taking advantage of its bifurcation structure by applying dynamic feedback control with a washout filter of ACI or IL-2. Bifurcation control moves the Hopf bifurcation point without changing the equilibrium structure as the bifurcation parameter varies. In this tumor-immune case, the linear dynamic feedback control with a washout filter of ACI could either extend the “safe equilibrium” region or reduce the amplitude of the tumor population at the stage of tumor recurrence. In addition, other bifurcation amplitude controls of either ACI or IL-2 are attempted to reduce the amplitudes of periodic orbits of the tumor immune system but without obvious effects.

Cavitation in cone valves can cause eventual valve failure. However, little data are available on cavitation limits for solid plug cone valves. Flow conditions were identified for a 6 inch cone valve where incipient, critical, incipient damage, and choking cavitation limits occurred at valve openings between 20 and 70 degrees. A description of how these limits are used to predict cavitation intensity is given. Air was used to reduce the damage levels. Aeration increased the allowable pressure drop for damage free operation 6 to 72 percent for valve openings less than 70 degrees. A chart is given for predicting when a cone valve will draw air. Location of the air injection point 1s not important as long as it is introduced into the separation zone.

abstract: In the last few decades, with the revolution of availability of low-cost microelectronics, which allow fast and complex computations to be performed on board, there has been increasing attention to aerial vehicles, especially rotary-wing vehicles. This is because of their ability to vertically takeoff and land (VTOL), which make them appropriate for urban environments where no runways are needed. Quadrotors took considerable attention in research and development due to their symmetric body, which makes them simpler to model and control compared to other configurations. One contribution of this work is the design of a new open-source based Quadrotor platform for research. This platform is compatible with both HTC Vive Tracking System (HVTS) and OptiTrack Motion Capture System, Robot Operating System (ROS), and MAVLINK communication protocol. The thesis examined both nonlinear and linear modeling of a 6-DOF rigid-body quadrotor's dynamics along with actuator dynamics. Nonlinear/linear models are used to develop control laws for both low-level and high-level hierarchical control structures. Both HVTS and OptiTrack were used to demonstrate path following for single and multiple quadrotors. Hardware and simulation data are compared. In short, this work establishes a foundation for future work on formation flight of multi-quadrotor. / Dissertation/Thesis / Masters Thesis Electrical Engineering 2019

abstract: Graphs are one of the key data structures for many real-world computing applica- tions such as machine learning, social networks, genomics etc. The main challenges of graph processing include diculty in parallelizing the workload that results in work- load imbalance, poor memory locality and very large number of memory accesses. This causes large-scale graph processing to be very expensive. This thesis presents implementation of a select set of graph kernels on a multi-core architecture, Transmuter. The kernels are Breadth-First Search (BFS), Page Rank (PR), and Single Source Shortest Path (SSSP). Transmuter is a multi-tiled architec- ture with 4 tiles and 16 general processing elements (GPE) per tile that supports a two level cache hierarchy. All graph processing kernels have been implemented on Transmuter using Gem5 architectural simulator. The key pre-processing steps in improving the performance are static partition- ing by destination and balancing the workload among the processing cores. Results obtained by processing graphs that are partitioned against un-partitioned graphs show almost 3x improvement in performance. Choice of data structure also plays an important role in the amount of storage space consumed and the amount of synchro- nization required in a parallel implementation. Here the compressed sparse column data format was used. BFS and SSSP are frontier-based algorithms where a frontier represents a subset of vertices that are active during the current iteration. They were implemented using the Boolean frontier array data structure. PR is an iterative algorithm where all vertices are active at all times. The performance of the dierent Transmuter implementations for the 14nm node were evaluated based on metrics such as power consumption (Watt), Giga Operations Per Second(GOPS), GOPS/Watt and L1/L2 cache misses. GOPS/W numbers for graphs with 10k nodes and 10k edges is 33 for BFS, 477 for PR and 10 for SSSP. i Frontier-based algorithms have much lower GOPS/W compared to iterative algo- rithms such as PR. This is because all nodes in Page Rank are active at all points in time. For all three kernel implementations, the L1 cache miss rates are quite low while the L2 cache hit rates are high. / Dissertation/Thesis / Masters Thesis Electrical Engineering 2019

abstract: Machine learning has demonstrated great potential across a wide range of applications such as computer vision, robotics, speech recognition, drug discovery, material science, and physics simulation. Despite its current success, however, there are still two major challenges for machine learning algorithms: limited robustness and generalizability. The robustness of a neural network is defined as the stability of the network output under small input perturbations. It has been shown that neural networks are very sensitive to input perturbations, and the prediction from convolutional neural networks can be totally different for input images that are visually indistinguishable to human eyes. Based on such property, hackers can reversely engineer the input to trick machine learning systems in targeted ways. These adversarial attacks have shown to be surprisingly effective, which has raised serious concerns over safety-critical applications like autonomous driving. In the meantime, many established defense mechanisms have shown to be vulnerable under more advanced attacks proposed later, and how to improve the robustness of neural networks is still an open question. The generalizability of neural networks refers to the ability of networks to perform well on unseen data rather than just the data that they were trained on. Neural networks often fail to carry out reliable generalizations when the testing data is of different distribution compared with the training one, which will make autonomous driving systems risky under new environment. The generalizability of neural networks can also be limited whenever there is a scarcity of training data, while it can be expensive to acquire large datasets either experimentally or numerically for engineering applications, such as material and chemical design. In this dissertation, we are thus motivated to improve the robustness and generalizability of neural networks. Firstly, unlike traditional bottom-up classifiers, we use a pre-trained generative model to perform top-down reasoning and infer the label information. The proposed generative classifier has shown to be promising in handling input distribution shifts. Secondly, we focus on improving the network robustness and propose an extension to adversarial training by considering the transformation invariance. Proposed method improves the robustness over state-of-the-art methods by 2.5% on MNIST and 3.7% on CIFAR-10. Thirdly, we focus on designing networks that generalize well at predicting physics response. Our physics prior knowledge is used to guide the designing of the network architecture, which enables efficient learning and inference. Proposed network is able to generalize well even when it is trained with a single image pair. / Dissertation/Thesis / Doctoral Dissertation Aerospace Engineering 2019