abstract: From 2D planar MOSFET to 3D FinFET, the geometry of semiconductor devices is getting more and more complex. Correspondingly, the number of mesh grid points increases largely to maintain the accuracy of carrier transport and heat transfer simulations. By substituting the conventional uniform mesh with non-uniform mesh, one can reduce the number of grid points. However, the problem of how to solve governing equations on non-uniform mesh is then imposed to the numerical solver. Moreover, if a device simulator is integrated into a multi-scale simulator, the problem size will be further increased. Consequently, there exist two challenges for the current numerical solver. One is to increase the functionality to accommodate non-uniform mesh. The other is to solve governing physical equations fast and accurately on a large number of mesh grid points. This research rst discusses a 2D planar MOSFET simulator and its numerical solver, pointing out its performance limit. By analyzing the algorithm complexity, Multigrid method is proposed to replace conventional Successive-Over-Relaxation method in a numerical solver. A variety of Multigrid methods (standard Multigrid, Algebraic Multigrid, Full Approximation Scheme, and Full Multigrid) are discussed and implemented. Their properties are examined through a set of numerical experiments. Finally, Algebraic Multigrid, Full Approximation Scheme and Full Multigrid are integrated into one advanced numerical solver based on the exact requirements of a semiconductor device simulator. A 2D MOSFET device is used to benchmark the performance, showing that the advanced Multigrid method has higher speed, accuracy and robustness. / Dissertation/Thesis / Masters Thesis Materials Science and Engineering 2015

Problemas de interface do tipo elípticos são frequentemente encontrados em dinâmicas de fluidos, ciências dos materiais, mecânica e outros campos de estudo. Em particular, o clássico Método de Interface Imersa (IIM) figura como uma das abordagens numéricas mais robustas para resolver problemas dessa categoria, o qual tem sido empregado recorrentemente para simular o comportamento de fluxos sobre corpos imersos em malhas cartesianas. Embora esse método seja eficiente e robusto, técnicas construídas com base no IIM impõem como restrições matemáticas diversos tipos de condições de salto na interface a fim de serem passíveis de utilização na prática. Nesta tese, introduzimos um novo método de Interface Imersa para resolver problemas elípticos com coeficientes descontínuos em malhas cartesianas. Diferentemente da maioria das formulações existentes que dependem de vários tipos de condições de salto para produzirem uma solução para o problema elíptico, o esquema aqui proposto reduz significativamente o número de restrições ao solucionar a EDP estudada, isto é, apenas os saltos de ordem zero das incógnitas devem ser fornecidos. A técnica apresentada combina esquemas de Diferenças Finitas, abordagem do Ponto Fantasma, modelos de correções e regras de interpolação em uma metodologia única e concisa. Além disso, o método proposto é capaz de produzir soluções de alta ordem, incluindo cenários onde há poucos dados disponíveis onde o quesito alta precisão é indispensável. A robustez e a precisão do método proposto são verificadas através de uma variedade de experimentos numéricos envolvendo diversos problemas elípticos com interfaces arbitrárias. Finalmente, a partir dos testes numéricos conduzidos, é possível concluir que o método projetado produz aproximações de alta ordem a partir de um número muito condensado de restrições matemáticas. / Elliptic interface problems are often encountered in fluid dynamics, material sciences, mechanics and other relevant fields of study. In particular, the well-known Immersed Interface Method (IIM) figures among the most effective approaches for solving non-trivial problems, where the method is traditionally used to simulate the flow behavior over complex bodies immersed in a cartesian mesh. Although their powerfulness and versatility, techniques that are built in light of the IIM impose as constraints different types of jump conditions at the interface in order to be properly managed and applicable for specific purposes. In this thesis, we introduce a novel Immersed Interface Method for solving Elliptic problems with discontinuous coefficients on cartesian grids. Different from most existing formulations that rely on various jump conditions types to get a valid solution, the present scheme reduces significatively the number of constraints when solving the PDE problem, i.e., only the ordinary jumps of the unknowns are required to be given, a priori. Our technique combines Finite Difference schemes, Ghost node strategy, correction models, and interpolation rules into a unified and concise methodology. Moreover, the method is capable of producing high-order solutions, succeeding in many practical scenarios with little available data wherein high precision is indispensable. We attest the robustness and the accuracy of the proposed method through a variety of numerical experiments involving several Elliptic problems with arbitrary interfaces. Finally, from the conducted numerical tests, we verify that the designed method produces high-order approximations from a very limited number of valid jump constraints.

The interdisciplinary field of machine learning studies algorithms in which functionality is dependent on data sets. This data is often treated as a matrix, and a variety of mathematical methods have been developed to glean information from this data structure such as matrix decomposition. The Laplacian matrix, for example, is commonly used to reconstruct networks, and the eigenpairs of this matrix are used in matrix decomposition. Moreover, concepts such as SVD matrix factorization are closely connected to manifold learning, a subfield of machine learning that assumes the observed data lie on a low-dimensional manifold embedded in a higher-dimensional space. Since many data sets have natural higher dimensions, tensor methods are being developed to deal with big data more efficiently. This thesis builds on these ideas by exploring how matrix methods can be extended to data presented as tensors rather than simply as ordinary vectors.

The aim of the thesis is to examine plasma-wall interaction using computer modeling. Tokamak- relevant plasma conditions are simulated using the particle-in-cell model family SPICE working in three or two dimensions. SPICE model was upgraded with a parallel Poisson equation solver and a heat equation solver module. Plasma simulation aimed at synthetic Langmuir probe measurements were performed. First set considered a flush-mounted probe and the effect of variable magnetic field angle was studied with aim to compare existing probe data evaluation techniques and assess their operational space, in which the plasma parameters estimation via fit to the current-voltage characteristic is accurate. Second simulation set studied a protruding probe pin. Effective collecting area of such probe was investigated with intentions of density measurement collection. This area was found to be influenced by a combination of two factors. First, the density dampening inside the magnetic pre-sheath of the probe head, and the second, the extension of the area caused by Larmor rotation. A comparison with experimental results obtained at COMPASS tokamak was was performed, confirming these results. Keywords Langmuir probe, simulation, particle-in-cell, tokamak, Poisson equation, COMPASS 1

Turbulent boundary layer and metamaterial properties were explored to initiate the viability of controlling acoustic waves driven by pressure fluctuations from flow. A turbulent boundary layer scaling analysis was performed on zero-pressure-gradient turbulent boundary layers over rough surfaces, for 30,000≤〖Re〗_θ≤100,000. Relationships between fluctuating pressures and velocities were explored through the pressure Poisson equation. Certain scaling laws were implemented in attempts to collapse velocity spectra and turbulence profiles. Such analyses were performed to justify a proper scaling of the low-frequency region of the wall-pressure spectrum. Such frequencies are commonly associated with eddies containing the largest length scales. This study compared three scaling methods proposed in literature: The low-frequency classical scaling (velocity scale U_τ, length scale δ), the convection velocity scaling (U_e-U ̅_c, δ), and the Zagarola-Smits scaling (U_e-U ̅, δ). A default scaling (U_e, δ) was also selected as a baseline case for comparison. At some level, the classical scaling best collapsed rough and smooth wall Reynolds stress profiles. Low-pass filtering of the scaled turbulence profiles improved the rough-wall scaling of the Zagarola-Smits and convection velocity laws. However, inconsistent scaled results between the pressure and velocity requires a more rigorous pressure Poisson analysis. The selection of a proper scaling law gives insight into turbulent boundary layers as possible sources for acoustic metamaterials. A quiescent (no flow) experiment was conducted to measure the capabilities of a metamaterial in retaining acoustic surface waves. A point source speaker provided an acoustic input while the resulting sound waves were measured with a probe microphone. Acoustic surface waves were found via Fourier analysis in time and space. Standing acoustic surface waves were identified. Membrane response properties were measured to obtain source condition characteristics for turbulent boundary layers once the metamaterial is exposed to flow. / Master of Science / Aerodynamicists are often concerned with interactions between fluids and solids, such as an aircraft wing gliding through air. Due to frictional effects, the relative velocity of the air on the solid-surface is negligible. This results in a layer of slower moving fluid near the surface referred to as a boundary layer. Boundary layers regularly occur in the fluid-solid interface, and account for a sufficient amount of noise and drag on aircraft. To compensate for increases in drag, engines are required to produce increased amounts of power. This leads to higher fuel consumption and increased costs. Additionally, most boundary layers in nature are turbulent, or chaotic. Therefore, it is difficult to predict the exact paths of air molecules as they travel within a boundary layer. Because of its intriguing physics and impacts on economic costs, turbulent boundary layers have been a popular research topic. This study analyzed air pressure and velocity measurements of turbulent boundary layers. Relationships between the two were drawn, which fostered a discussion of future works in the field. Mainly, the simultaneous measurements of pressure on the surface and boundary layer velocity can be performed with understanding of the Pressure Poisson equation. This equation is a mathematical representation of the boundary layer pressure on the surface. This study also explored the possibility of turbulent-boundary-layer-driven-acoustic-metamaterials. Acoustic metamaterials contain hundreds of cavities which can collectively manipulate passing sound waves. A facility was developed at Virginia Tech to measure this effect, with aid from a similar laboratory at Exeter University. Microphone measurements showed the reduction of sound wave speed across the metamaterial, showing promise in acoustic manipulation. Applications in metamaterials in the altering of sound caused by turbulent boundary layers were also explored and discussed.

Made available in DSpace on 2016-12-23T14:08:15Z (GMT). No. of bitstreams: 1 Atila Lupim Cruz.pdf: 1394501 bytes, checksum: 0954b2c5b1fdcb864ee81cef7d14e9e5 (MD5) Previous issue date: 2012-10-19 / A pesquisa envolvida na presente dissertação se baseou no uso de funções de base radial para gerar uma nova formulação integral, que interpola diretamente o termo não homogêneo da equação diferencial de governo, no contexto do Método dos Elementos de Contorno (MEC). Emprega-se o uso de funções primitivas das funções de interpolação originais no núcleo da integral de domínio, permitindo a transformação desta última numa integral de contorno, evitando assim a discretização do domínio por meio de células, semelhante ao realizado na Dupla Reciprocidade. Para melhor avaliação das potencialidades da formulação, os testes numéricos apresentados abordaram apenas a solução de problemas governados pela Equação de Poisson. Os problemas escolhidos dentro desta categoria possuem solução analítica, o que permitiu aferir com mais rigor a precisão dos resultados. Para melhor balizamento da eficiência da formulação proposta, todos os problemas abordados também foram resolvidos pela formulação com Dupla Reciprocidade. O custo computacional dispendido para cada uma dessas formulações também foi comparado. Para ambas as formulações também foram testados esquemas de ajuste da interpolação realizada, visando avaliar seus efeitos na precisão dos resultados e também propositando obter economia computacional em futuras aplicações em simulações na área de propagações de ondas / This research was based on the use of radial basis functions to generate a new integral formulation that interpolates directly the domain action, related to the inhomogeneous term of the governing differential equation, using the Boundary Element Method (BEM). The use of primitive functions of the original interpolation functions in the kernel of the inhomogeneous integral is proposed, allowing its transformation into a boundary integral, thus avoiding the domain discretization through cells, similar to that conducted in the Dual Reciprocity. To better evaluation of the capability of the proposed formulation, the numerical tests presented only solved problems governed by the Poisson Equation. Test problems chosen have known analytical solution, which allowed a better evaluation of the numerical accuracy. To better check the efficiency of the proposed formulation, all the problems were also solved by the Dual Reciprocity Boundary Element Formulation. The computational cost expended for each of these formulations was also compared. Fitting interpolation schemes for both formulations were also tested in order to evaluate their effects on the accuracy of the results and also looking for economy in future computational applications related to wave propagation problems

Various properties of linear infinite-dimensional Hamiltonian systems are studied. The structural stability of the Vlasov-Poisson equation linearized around a homogeneous stable equilibrium [mathematical symbol] is investigated in a Banach space setting. It is found that when perturbations of [mathematical symbols] are allowed to live in the space [mathematical symbols], every equilibrium is structurally unstable. When perturbations are restricted to area preserving rearrangements of [mathematical symbol], structural stability exists if and only if there is negative signature in the continuous spectrum. This analogizes Krein's theorem for linear finite-dimensional Hamiltonian systems. The techniques used to prove this theorem are applied to other aspects of the linearized Vlasov-Poisson equation, in particular the energy of discrete modes which are embedded within the continuous spectrum. In the second part, an integral transformation that exactly diagonalizes the Caldeira-Leggett model is presented. The resulting form of the Hamiltonian, derived using canonical transformations, is shown to be identical to that of the linearized Vlasov-Poisson equation. The damping mechanism in the Caldeira-Leggett model is identified with the Landau damping of a plasma. The correspondence between the two systems suggests the presence of an echo effect in the Caldeira-Leggett model. Generalizations of the Caldeira-Leggett model with negative energy are studied and interpreted in the context of Krein's theorem. / text

Nowadays there is an ever increasing demand to obtain more accurate numericalsimulation results while at the same time using fewer computations. One area withsuch a demand is oil reservoir simulations, which builds upon Poisson's equation withvariable coefficients (PEWVC). This thesis focuses on applying and testing a high ordernumerical scheme to solve the PEWVC, namely Summation By Parts - SimultaneousApproximation Term (SBP-SAT). The thesis opens with proving that the method isconvergent at arbitrary high orders given sufficiently smooth coefficients. Theconvergence is furthermore verified in practice by test cases on the Poisson'sequation with smoothly variable permeability coefficients. To balance observed lowerboundary flux convergence, the SBP-SAT method was modified with additionalpenalty terms that were subsequently shown to work as expected. Finally theSBP-SAT method was tested on a semi-realistic model of an oil reservoir withdiscontinuous permeability. The correctness of the resulting pressure distributionvaried and it was shown that flux leakage was the probable cause. Hence theproposed SBP-SAT method performs, as expected, very well in continuous settingsbut typically allows undesirable leakage in discontinuous settings. There are possiblefixes, but these are outside the scope of this thesis.

PINTO, V. G. Caracterizações da esfera em formas espaciais. 2017. 79 f. Dissertação (Mestrado em Matemática) – Centro de Ciências, Universidade Federal do Ceará, Fortaleza, 2017. / Submitted by Andrea Dantas (pgmat@mat.ufc.br) on 2017-07-20T20:40:07Z No. of bitstreams: 1 2017_dis_vgpinto.pdf: 1180135 bytes, checksum: f3aa196ed8b0d38c5a2a33642fdb7d0b (MD5) / Rejected by Rocilda Sales (rocilda@ufc.br), reason: Bom dia Andrea, Favor informar ao aluno os motivos da rejeição. Faltou a conclusão (item obrigatório) E as referências não estão normalizadas. Seguem os modelos ARTIGOS DE PERIÓDICOS: ALENCAR, H. ; COLARES, A. G. - Integral formulas for the r-mean curvature linearized operator of a hypersurface. Annals of Global Analysis and Geometry, v. 16, p. 203-220, 1998. OBS: o TÍTULO DO PERIÓDICO DEVE FICAR EM NEGRITO OU ITÁLICO. LIVROS: CARMO, M. P. do. Geometria riemanniana. Rio de Janeiro : IMPA, 2008.( Projeto Euclides) OBS: O TÍTULO DO LIVRO DEVE FICAR EM NEGRITO OU ITÁLICO DISSERTAÇÕES: PINHEIRO, N. R. Hipersuperfíıcies com curvatura média constante e hiperplanos. Ano. Nº de folhas. Dissertação ( Mestrado) em nome do curso, local, ano. OBS: o TÍTULO DA DISSERTAÇÃO DEVE FICAR EM NEGRITO OU ITÁLICO Rocilda on 2017-07-21T11:38:59Z (GMT) / Submitted by Andrea Dantas (pgmat@mat.ufc.br) on 2017-07-21T18:48:58Z No. of bitstreams: 1 2017_dis_vgpinto.pdf: 1184804 bytes, checksum: 357d2ee050e65edb2839093ba455b0db (MD5) / Approved for entry into archive by Rocilda Sales (rocilda@ufc.br) on 2017-07-24T15:34:13Z (GMT) No. of bitstreams: 1 2017_dis_vgpinto.pdf: 1184804 bytes, checksum: 357d2ee050e65edb2839093ba455b0db (MD5) / Made available in DSpace on 2017-07-24T15:34:13Z (GMT). No. of bitstreams: 1 2017_dis_vgpinto.pdf: 1184804 bytes, checksum: 357d2ee050e65edb2839093ba455b0db (MD5) Previous issue date: 2017-07-06 / In this work we present three characterizations of the sphere. Initially, it will be shown that given a compact and oriented hypersurface Mn e x: M → Q^(n+1)_c a isometric immersion, x(M) is a geodesic sphere in Q^n+1_c if, and only if, Hr+1 is a nonzero constant and the set of points that are omitted in Qn+1 c by the totally geodesic hypersurfaces (Q^n_c)p tangent to x(M) is non-empty. As a second result, let M be an orientable compact and connected hypersurface with non-negative support function of the Euclidean space Rn+1 and Minkowski's integrand . We prove that the mean curvature function of the hypersurface M is the solution of the Poisson equation = if, and only if, M is isometric to the n-sphere Sn(c) of constant curvature c. similar characterization is proved for a hypersurface with the scalar curvature satisfying the same equation. For the third result we consider an isometric immersion x : M ! Qn+1, where M is a compact hypersurface such that x(M) is convex, and it will be proved that if any r-mean curvature is such that Hr 6= 0 and there are nonnegative constants C1;C2; :::;Cr1 such that Hr = Pr1 i=1 CiHi; then x(M) is a geodesic sphere, where Qn+1 is Rn+1, Hn+1 or Sn+1 + . / Neste trabalho serão apresentadas três caracterizações da esfera. Primeiramente, será mostrado que dada uma hipersuperfície compacta e orientada Mn e x: M → Q^(n+1)_c uma imersão isométrica, onde Q^n+1_c é uma forma espacial simplesmente conexa, isto é, uma variedade Riemanniana de curvatura seccional constante c, x(M) é uma esfera geodésica em Q^n+1_c se, e somente se, a (r + 1)-ésima curvatura média Hr+1 é uma constante não nula e o conjunto dos pontos que são omitidos em Q^n+1_c pelas hipersuperfícies totalmente geodésicas (Q^n_c)p tangentes a x(M) é não vazio. Como segundo resultado, seja uma hipersuperfície compacta, conexa e orientável M do espaço euclidiano R^(n+1), com função suporte não negativa e integrando de Minkowski σ. Será provado que a função curvatura média α da hipersuperfície é solução da equação de Poisson Δϕ = σ se, e somente se, M é isométrica à n-esfera S^n(c) de curvatura média c. Uma caracterização similar é provada para uma hipersuperfície com a curvatura escalar satisfazendo a mesma equação. Para o terceiro resultado é considerado uma imersão isométrica x: M → Q^(n+1), onde M é uma hipersuperfície compacta tal que x(M) é convexa, e será provado que, se alguma curvatura r-média é tal que Hr ≠ 0 e existem constantes não negativas C1, C2, ..., Cr-1 tais que Hr =∑_(i=1)^(r-1)▒〖C_i H_i 〗 ; então x(M) é uma esfera geodésica, onde Q^(n+1) é R^(n+1), H^(n+1) ou S^(n+1)_+ .

Projection methods for the incompressible Navier-Stokes equations (NSE) are efficient, but introduce numerical boundary layers and have limited temporal accuracy due to their fractional step nature. The Pressure Poisson Equation (PPE) reformulations represent a class of methods that replace the incompressibility constraint by a Poisson equation for the pressure, with a suitable choice of the boundary condition so that the incompressibility is maintained. PPE reformulations of the NSE have important advantages: the pressure is no longer implicitly coupled to the velocity, thus can be directly recovered by solving a Poisson equation, and no numerical boundary layers are generated; arbitrary order time-stepping schemes can be used to achieve high order accuracy in time. In this thesis, we focus on numerical approaches of the PPE reformulations, in particular, the Shirokoff-Rosales (SR) PPE reformulation. Interestingly, the electric boundary conditions, i.e., the tangential and divergence boundary conditions, provided for the velocity in the SR PPE reformulation render classical nodal finite elements non-convergent. We propose two alternative methodologies, mixed finite element methods and meshfree finite differences, and demonstrate that these approaches allow for arbitrary order of accuracy both in space and in time. / Mathematics