O presente trabalho teve como objetivos: (1) avaliar o rendimento e as características físico-químicas dos óleos de vísceras de truta arco-íris (Oncorhynchus mykiss), pacu (Piaractus mesopotamicus) e curimbatá (Prochilodus spp.) extraídos por congelamento lento, como mecanismo para liberar os lipídios do material bruto, e (2) comparar as etapas do refinamento químico (degomagem, neutralização, lavagem, secagem e branqueamento) de dois óleos de truta arco-íris, extraídos por congelamento lento e por um processo termo-mecânico. As matérias primas foram coletadas em diferentes municípios dos Estados de São Paulo e Paraná. Foram determinados os rendimentos, teor de ácidos graxos livres, índice de peróxidos, índice de iodo, índice de saponificação, densidade, índice de refração, viscosidade e perfil de ácidos graxos nos óleos brutos (truta arco-íris, pacu e curimbatá) e nos óleos submetidos a processos de refinamento (truta arco-íris). Os rendimentos de óleo de vísceras de truta arco-íris, pacu e curimbatá, foram de 27,58%, 42,53 e 13,75% respectivamente. Os parâmetros físicoquímicos avaliados em todos os óleos (óleos brutos das vísceras das três espécies de peixes e óleos de truta arco-íris submetidos a refinamento) encontraram-se dentro dos níveis referenciados para óleos de peixe brutos e refinados. O perfil de ácidos graxos foi variável entre os óleos brutos das três espécies, sendo que o óleo de curimbatá apresentou maiores níveis dos ácidos eicosapentaenoico, docosaexaenoico e araquidônico, e menor relação n-6/n-3, portanto de melhor qualidade nutricional, provavelmente por ter sido capturado na natureza. Entretanto, devido à alta disponibilidade de truta arco-íris, considera-se mais viável a produção de complementos nutricionais de ácidos graxos poliinsaturados da família n-3 nesta espécie, desde que o perfil lipídico do óleo esteja adequado. Os teores de ácidos graxos da família n-3, dos óleos de truta arco-íris foram menores que os relatados na literatura; podendo ser influenciados pela dieta dos animais. O tipo de extração produziu diferenças no perfil de ácidos graxos tanto nos óleos brutos quanto nos óleos branqueados. Destaca-se maior teor de ácidos graxos com configuração trans no óleo (bruto e branqueado) de truta arco-íris, extraído por autoclavagem devido provavelmente à utilização de altas temperaturas durante a extração. A técnica de extração por congelamento foi eficiente em termos de rendimento e qualidade dos óleos extraídos das vísceras dos peixes avaliados, e possivelmente seja eficiente também em outras espécies e matérias primas. / This study aimed to: (1) evaluate the performance and physicochemical characteristics of viscera oils from rainbow trout (Oncorhynchus mykiss), pacu (Piaractus mesopotamicus) and curimbatá (Prochilodus spp.) extracted by slow freezing, as a mechanism to release the lipids of the crude material, and (2) compare the stages of chemical refining (degumming, neutralization, washing, drying and bleaching) of two rainbow trout oils, extracted by slow freezing and by a thermo-mechanic process. The raw materials were collected in different places of the States of São Paulo and Parana. Viscera oil yields were determined, content of free fatty acids, peroxide value, iodine value, saponification value, density, refractive index, viscosity and fatty acid profile of the crude oils (rainbow trout, pacu and curimbatá) and of the rainbow trout oils underwent the refinement process. The oil yields of viscera rainbow trout, pacu and curimbatá, were 27.58%, 42.53 and 13.75% respectively. The physico-chemical parameters evaluated in all the oils (crude oils from the viscera of the three species of fish and rainbow trout oils subjected to refinement) were within the referenced levels for the crude and refined fish oils. The fatty acid profile varied between the crude oils from the three species. Curimbatá viscera oil, showed higher levels of eicosapentaenoic acid, docosahexaenoic and arachidonic acid, and lower n-6/n-3 ratio therefore of higher nutritional quality, probably for having been caught in the wild. However, due to the high availability of rainbow trout, it is more viable the production of nutritional supplements of n-3 polyunsaturated fatty acids with by-products from this species, since their oil lipid profile is suitable. The levels of n-3 fatty acids of rainbow trout oils were lower than those reported in the literaturepossibly influenced by the diet. The type of extraction produced differences in the fatty acid profile both in crude oils and bleached oils. It stands out higher content of fatty acids with trans configuration in the rainbow trout oil (crude and bleached), obtained by autoclaving probably due to the use of high temperatures during extraction. The freeze-extraction technique was efficient in terms of yield and quality of oils extracted from the viscera of fishes evaluated and could also be effective in other species and raw materials.

General relativity, one of the pillars of our understanding of the universe, has been a remarkably successful theory. It has stood the test of time for more than 100 years and has passed all experimental tests so far. Most recently, the LIGO collaboration made the first-ever direct detection of gravitational waves, confirming a long-standing prediction of general relativity. Despite this, several fundamental mathematical questions remain unanswered, many of which relate to the global existence and the stability of solutions to Einstein's equations. This thesis presents our efforts to use numerical relativity to investigate some of these questions. We present a complete picture of the end points of black ring instabilities in five dimensions. Fat rings collapse to Myers-Perry black holes. For intermediate rings, we discover a previously unknown instability that stretches the ring without changing its thickness and causes it to collapse to a Myers-Perry black hole. Most importantly, however, we find that for very thin rings, the Gregory-Laflamme instability dominates and causes the ring to break. This provides the first concrete evidence that in higher dimensions, the weak cosmic censorship conjecture may be violated even in asymptotically flat spacetimes. For Myers-Perry black holes, we investigate instabilities in five and six dimensions. In six dimensions, we demonstrate that both axisymmetric and non-axisymmetric instabilities can cause the black hole to pinch off, and we study the approach to the naked singularity in detail. Another question that has attracted intense interest recently is the instability of anti-de Sitter space. In this thesis, we explore how breaking spherical symmetry in gravitational collapse in anti-de Sitter space affects black hole formation. These findings were made possible by our new open source general relativity code, GRChombo, whose adaptive mesh capabilities allow accurate simulations of phenomena in which new length scales are produced dynamically. In this thesis, we describe GRChombo in detail, and analyse its performance on the latest supercomputers. Furthermore, we outline numerical advances that were necessary for simulating higher dimensional black holes stably and efficiently.

We present INCREMENT, a cluster refinement algorithm which utilizes user feedback to refine clusterings. INCREMENT is capable of improving clusterings produced by arbitrary clustering algorithms. The initial clustering provided is first sub-clustered to improve query efficiency. A small set of select instances from each of these sub-clusters are presented to a user for labelling. Utilizing the user feedback, INCREMENT trains a feature embedder to map the input features to a new feature space. This space is learned such that spatial distance is inversely correlated with semantic similarity, determined from the user feedback. A final clustering is then formed in the embedded space. INCREMENT is tested on 9 datasets initially clustered with 4 distinct clustering algorithms. INCREMENT improved the accuracy of 71% of the initial clusterings with respect to a target clustering. For all the experiments the median percent improvement is 27.3% for V-Measure and is 6.08% for accuracy.

Proteins are the fundamental machinery that enables the functions of life. It is critical to understand them not just for basic biology, but also to enable medical advances. The field of protein structure prediction is concerned with developing computational techniques to predict protein structure and function from a protein’s amino acid sequence, encoded for directly in DNA, alone. Despite much progress since the first computational models in the late 1960’s, techniques for the prediction of protein structure still cannot reliably produce structures of high enough accuracy to enable desired applications such as rational drug design. Protein structure refinement is the process of modifying a predicted model of a protein to bring it closer to its native state. In this dissertation a protein structure refinement technique, that of potential energy minimization using hybrid molecular mechanics/knowledge based potential energy functions is examined in detail. The generation of the knowledge-based component is critically analyzed, and in the end, a potential that is a modest improvement over the original is presented. This dissertation also examines the task of protein structure comparison. In evaluating various protein structure prediction techniques, it is crucial to be able to compare produced models against known structures to understand how well the technique performs. A novel technique is proposed that allows an in-depth yet intuitive evaluation of the local similarities between protein structures. Based on a graph analysis of pairwise atomic distance similarities, multiple regions of structural similarity can be identified between structures independently of relative orientation. Multidomain structures can be evaluated and this technique can be combined with global measures of similarity such as the global distance test. This method of comparison is expected to have broad applications in rational drug design, the evolutionary study of protein structures, and in the analysis of the protein structure prediction effort.

This thesis provides an in depth look at the application of Shot peening to three alloys that cover a wide range of stabilizer compositions (Ti-64, Ti-5553 and Ti21S). Here the effects of Shot peening on the material were studied by characterizing the changes to subsurface microstructres in a variety of starting microstructures across all alloys studied. the study also compared the evolution of residual stress across all conditions and its effect on precipitation during post shot peening heat treatments. <div>Additionally, the study developed a novel shot peening media providing proof of concept for the idea of self-peening a material. This being the practice of peening the target material with the same material as media. Here the production of hardened Ti media was achieved to then self peen titanium alloys. the results from conventional and self-peened samples are compared in terms of corrosion resistance, surface hardening, surface roughness, and residual stress evolution.</div>

A major drawback in the operation of mechanical heart valve prostheses is thrombus formation in the near valve region potentially due to the high shear stresses present in the leakage jet flows through small gaps between leaflets and the valve housing. Detailed flow analysis in this region during the valve closure phase is of interest in understanding the relationship between shear stress and platelet activation. An efficient Cartesian grid method is developed for the simulation of incompressible flows around stationary and moving three-dimensional immersed solid bodies as well as fluid-fluid interfaces. The embedded boundaries are represented using Levelsets and treated in a sharp manner without the use of source terms to represent boundary effects. The resulting algorithm is implemented in a straightforward manner in three dimensions and retains global second-order accuracy. When dealing with problems of disparate length scales encountered in many applications, it is necessary to resolve the physically important length scales adequately to ensure accuracy of the solution. Fixed grid methods often have the disadvantage of heavy mesh requirement for well resolved calculations. A quadtree based adaptive local mesh refinement scheme is developed to complement the sharp interface Cartesian grid method scheme for efficient and optimized calculations. Detailed timing and accuracy data is presented for a variety of benchmark problems involving moving boundaries. The above method is then applied to modeling heart valve closure and predicting thrombus formation. Leaflet motion is calculated dynamically based on the fluid forces acting on it employing a fluid-structure interaction algorithm. Platelets are modeled and tracked as point particles by a Lagrangian particle tracking method which incorporates the hemodynamic forces on the particles. Leaflet closure dynamics including rebound is analyzed and validated against previous studies. Vortex shedding and formation of recirculation regions are observed downstream of the valve, particularly in the gap between the valve and the housing. Particle exposure to high shear and entrapment in recirculation regions with high residence time in the vicinity of the valve are observed corresponding to regions prone to thrombus formation.

In this thesis a method for direct evaluation of Generalized B-splines (GB-splines) via the representation of these curves as piecewise functions is presented. A local structure is introduced that makes the GB-spline curves more amenable to the integration used in constructing bases of higher degree. This basis is used to perform direct computation of piecewise representation of GB-spline bases and curves. Algorithms for refinement using these local structures are also developed.

β-grain size is an influential microstructural parameter on the properties of titanium components. A reduction of β-grain size is generally associated with improvements to ductility, strength, corrosion and fatigue resistance of many α, α/β and β titanium alloys. During production of wrought titanium components, the β-grain size is carefully controlled during thermomechanical processing but there is currently no control of the β-grain size during solidification of cast components. As such, this inability to control the β-grain structure during solidification may limit the applications for solidification based technologies including casting, welding and direct metal deposition. Due to the limited knowledge of grain refinement practices and the lack of commercial grain refiners for the titanium system, this thesis investigates the mechanisms of β-grain refinement during solidification of cast titanium alloys. In this thesis, generalized theories for grain refinement that have been developed from research into other metallic systems are applied to the titanium system. Similar to the findings from aluminium and magnesium research, it is shown that grain refinement of cast titanium alloys requires the addition of growth restricting solutes which provide constitutional undercooling as well as the presence of potent nucleant particles. It is demonstrated that commercially pure titanium contains a natural distribution of nuclei particles which may originate from the mould wall and when powerful growth restricting solutes are introduced, significant prior-β grain refinement is achievable. All solutes investigated do not interact or poison the naturally occurring nucleants enabling the grain size of the titanium alloys to be predicted by an empirically determined relationship based on the growth restriction factor. A full list of growth restriction factors for various elements in titanium is determined and it is proven that growth restriction theory is valid in the titanium system. A further reduction in β-grain size is achievable by introducing additional nucleant particles to titanium castings in conjunction with growth-restricting solutes. Using a novel technique, titanium powder was introduced to the melt stream prior to solidification and was mixed throughout the liquid. The powder particles partially melted and the oxide surface layer dissolved allowing intimate substrate-liquid contact, enabling the titanium substrates to act as sites for heterogeneous nucleation. Using this technique, it was possible to grain refine commercially pure titanium without foreign elemental addition and when growth restricting solutes were present it was possible to obtain approximately an order of magnitude grain size reduction. The results and concepts developed from this work may aid the future development of a commercial grain refiner for titanium. If a grain refiner is developed, its application will not just be limited to the titanium casting industry but may also benefit other solidification based technologies such as welding, direct metal deposition and wrought billet production.

Concurrency is concerned with systems of multiple computing agents that interact with each other. Bisimilarity is one of the main representatives of these. Concurrent Constrain Programming (ccp) is a formalism that combines the traditional and algebraic view of process calculi with a declarative one based upon first-order logic. The standard definition of bisimilarity is not completely satisfactory for ccp since it yields an equivalence that is too fine grained. By building upon recent foundational investigations, we introduce a labeled transition semantics and a novel notion of bisimilarity that is fully abstract w.r.t. the observational equivalence in ccp. When the state space of a system is finite, the ordinary notion of bisimilarity can be computed via the partition refinement algorithm, but unfortunately, this algorithm does not work for ccp bisimilarity. Hence, we provide an algorithm that allows us to verify strong bisimilarity for ccp, modifying the algorithm by using a pre-refinement and a partition function based on the irredundant bisimilarity. Weak bisimilarity is a central behavioral equivalence in process calculi and it is obtained from the strong case by taking into account only the actions that are observable in the system. Typically the standard partition refinement can also be used for deciding weak bisimilarity simply by using Milner's reduction from weak to strong; a technique referred to as saturation. We demonstrate that the above-mentioned saturation technique does not work for ccp. We give a reduction that allows us to use the ccp partition refinement algorithm for deciding this equivalence.

A spectral element algorithm for solution of the incompressible Navier-Stokes and heat transfer equations is developed, with an emphasis on extending the classical conforming Galerkin formulations to nonconforming spectral elements. The new algorithm employs both the Constrained Approximation Method (CAM), and the Mortar Element Method (MEM) for p-and h-type nonconforming elements. Detailed descriptions, and formulation steps for both methods, as well as the performance comparisons between CAM and MEM, are presented. This study fills an important gap in the literature by providing a detailed explanation for treatment of p-and h-type nonconforming interfaces. A comparative eigenvalue spectrum analysis of diffusion and convection operators is provided for CAM and MEM. Effects of consistency errors due to the nonconforming formulations on the convergence of steady and time dependent problems are studied in detail. Incompressible flow solvers that can utilize these nonconforming formulations on both p- and h-type nonconforming grids are developed and validated. Engineering use of the developed solvers are demonstrated by detailed parametric analyses of oscillatory flow forced convection heat transfer in two-dimensional channels.