This thesis deals with the problematic of test-data generation for relational databases. The aim of the this thesis is to design and implement tool which meets defined constrains and allows us to generate test-data. This tool uses SMT solver for constraint solving and test-data generation.

abstract: Semiconductor devices often face reliability issues due to their operational con- ditions causing performance degradation over time. One of the root causes of such degradation is due to point defect dynamics and time dependent changes in their chemical nature. Previously developed Unified Solver was successful in explaining the copper (Cu) metastability issues in cadmium telluride (CdTe) solar cells. The point defect formalism employed there could not be extended to chlorine or arsenic due to numerical instabilities with the dopant chemical reactions. To overcome these shortcomings, an advanced version of the Unified Solver called PVRD-FASP tool was developed. This dissertation presents details about PVRD-FASP tool, the theoretical framework for point defect chemical formalism, challenges faced with numerical al- gorithms, improvements for the user interface, application and/or validation of the tool with carefully chosen simulations, and open source availability of the tool for the scientific community. Treating point defects and charge carriers on an equal footing in the new formalism allows to incorporate chemical reaction rate term as generation-recombination(G-R) term in continuity equation. Due to the stiff differential equations involved, a reaction solver based on forward Euler method with Newton step is proposed in this work. The Jacobian required for Newton step is analytically calculated in an elegant way improving speed, stability and accuracy of the tool. A novel non-linear correction scheme is proposed and implemented to resolve charge conservation issue. The proposed formalism is validated in 0-D with time evolution of free carriers simulation and with doping limits of Cu in CdTe simulation. Excellent agreement of light JV curves calculated with PVRD-FASP and Silvaco Atlas tool for a 1-D CdTe solar cell validates reaction formalism and tool accuracy. A closer match with the Cu SIMS profiles of Cu activated CdTe samples at four different anneal recipes to the simulation results show practical applicability. A 1D simulation of full stack CdTe device with Cu activation at 350C 3min anneal recipe and light JV curve simulation demonstrates the tool capabilities in performing process and device simulations. CdTe device simulation for understanding differences between traps and recombination centers in grain boundaries demonstrate 2D capabilities. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2019

CFD-Analysen sind in den letzten Jahren ein fester Bestandteil in der Produktentwicklung geworden. Dies liegt primär daran, dass es in den letzten Jahrzehnten große Fortschritte in der Entwicklung der Software und dem Preisverfall der Hardware gegeben hat. Das einstige Expertenwerkzeug CFD hat heutzutage auch in vielen kleinen und mittelständischen Unternehmen seinen festen Platz gefunden und somit ein enormes Anwendungsspektrum erschlossen. Ein wesentlicher Schlüssel für diesen Erfolg war und ist die Effizienz bei der Erstellung der Berechnungsmodelle. Ein Meilenstein war die ab Mitte der 1990er Jahre verfügbare Verwendung unstrukturierter Netze. Erst durch den Einsatz von Tetraedern konnten beliebig komplexe Strukturen überhaupt vernetzt werden. Dieser Vortrag beschreibt die Verwendung von sogenannten kartesischen Netzen in Kombination mit der Immersed Boundary Methode. Dabei handelt es sich um einen bislang wenig beachteten Weg, der den Anwender nochmals erheblich bei der Modellaufbereitung entlastet und somit auch anspruchsvolle CFD-und Thermalsimulationen für Konstrukteure zugänglich macht.

<p>NR 20140805</p>

This thesis discusses the application, evaluation, and extension of dimensionally adaptive meshing to the numerical solution of velocity and pressure fields inside cylindrical reactors. Due to the high length to diameter ratios of many cylindrical reactor vessels the flow field can become axisymmetric, allowing for simplification of the governing equations and significant reduction in computational time required for solution. A fully 3D solver is developed from existing computational tools at BYU and validated against theoretical velocity profiles for pipe flow at various Reynolds numbers, as well as with experimental data for an axial-fired center jet with recirculating flow. Dimensionally adaptive meshing is then incorporated into the validated 3D solver. The boundary conditions and assumptions at the dimensional boundary are discussed. The flow information is passed across the boundary through spatial mass-weighted averaging. The 3D and axisymmetric computational domains are decoupled from one another so information can only be passed from the 3D domain downstream to the axisymmetric domain. The dimensional boundary placement must meet two main requirements, the flow must be one-way and axisymmetric. It is found that the flow becomes axisymmetric early on in the reactor (~0.3-0.4 m), but recirculation exists farther downstream (until ~0.61 m) and thus governs the placement of the dimensional boundary. The resulting computational tool capable of running simulations using dimensionally adaptive meshes is called CEDAR (Computationally Efficient Dimensionally Adaptive Recirculating flow solver). Several studies are then undertaken to examine CEDAR's ability to reproduce exit velocity profiles comparable to those produced by a fully 3D mesh, including variations in pressure, firing rate, and geometry. It is found that the flow structure inside the reactor is self-similar over a wide range of operating parameters as long as the burner jets are turbulent. This observation is supported by free and confined jet theory. These theories also provide a method for placing the dimensional boundary, which is a linear function of the confining geometry diameter only (assuming that the jet diameter is less than 1/10 the diameter of the confining geometry). All exit velocity profiles produced by CEDAR are on average within 5% of the fully 3D profiles. Timing studies reveal an average 5.16 times speedup in computational time over fully 3D computations.

Time-domain methods are preferred over their frequency-domain counterparts for solving acoustic and electromagnetic scattering problems since they can produce wide- band data from a single simulation. Among the time-domain methods, time-domain surface integral equation solvers have recently found widespread use because they offer several benefits over differential equation solvers. This dissertation develops several second-kind surface integral equation solvers for analyzing transient acoustic scattering from rigid and penetrable objects and transient electromagnetic scattering from perfect electrically conducting and dielectric objects. For acoustically rigid, perfect electrically conducting, and dielectric scatterers, fully explicit marching-on-in-time schemes are developed for solving time domain Kirchhoff, magnetic field, and scalar potential integral equations, respectively. The unknown quantity (e.g., velocity potential, electric current, or scalar potential) on the scatterer surface is discretized using a higher-order method in space and Lagrange interpolation in time. The resulting system is cast in the form of an ordinary differen- tial equation and integrated in time using a predictor-corrector scheme to obtain the unknown expansion coefficients. The explicit scheme can use the same time step size as its implicit counterpart without sacrificing from the stability of the solution and is much faster under low-frequency excitation (i.e., for large time step size). In addition, low-frequency behavior of vector potential integral equations for perfect electrically conducting scatterers is also investigated in this dissertation. For acoustically penetrable scatterers, presence of spurious interior resonance modes in the solutions of two forms of time domain surface integral equations is investigated. Numerical results demonstrate that the solution of the form that is widely used in the literature is corrupted by the interior resonance modes. But, the amplitude of these modes in the time domain can be suppressed by increasing the accuracy of discretization especially in time. On the other hand, the proposed one in the combined form shows a resonance-free performance verified via numerical experiments. In addition to providing detailed formulations of these solvers, the dissertation presents numerical examples, which demonstrate the solvers’ accuracy, efficiency, and applicability in real-life scenarios.

Dans ce manuscrit, nous présentons une méthode d’optimisation de forme qui se base sur des paramètres géométriques comme des longueurs, des angles, etc. Nous nous appuyons sur des techniques d’optimisation basées sur un gradient. La sensibilité de la fonction objectif par rapport à la position des noeuds du maillage nous est fournie par un solveur adjoint que l’on considère comme une boîte noire. Afin d’optimiser par rapport aux paramètres CAO, nous nous concentrons sur l’évaluation de la sensibilité de la position des noeuds par rapport à ces paramètres. Ainsi, nous proposons deux approches par différences finies. La première méthode s’appuie sur une projection harmonique afin de comparer dans un même espace le maillage initial et celui obtenu suite à la variation d’un paramètre CAO. Les développements présentés dans ce manuscrit permettent d’étendre l’application aux formes ayant plusieurs frontières comme les collecteurs d’échappement. Nous avons développé une méthode d’interpolation adaptée à cette comparaison. L’ensemble du processus a été automatisé et nous en montrons l’entière efficacité sur des applications industrielles en aérodynamique interne. La deuxième méthode se base directement sur les géométries CAO pour évaluer cette sensibilité. Nous utilisons la définition intrinsèque des patches dans l’espace paramétrique (u;v) pour effectuer cette comparaison. Grâce à l’utilisation des coordonnées exactes en tout point de la surface fournies par la CAO, nous évitons d’avoir recours à une interpolation afin d’avoir la meilleure précision de calcul possible. Cependant, contrairement à la première méthode, elle requiert d’identifier les correspondances entre les patches d’une forme à l’autre. Une application sur un cas académique a été faite en aérodynamique externe. La pertinence de la première méthode a été démontrée sur des cas représentatifs et multiobjectifs, ce qui permettrait de faciliter son déploiement et son utilisation dans un cadre industriel. Quant à la deuxième méthode, nous avons montré son fort potentiel. Cependant, des développements supplémentaires seraient nécessaires pour une application plus poussée. Du fait qu’elles sont indépendantes des solveurs mécaniques et du nombre de paramètres, ces méthodes réduisent considérablement les temps de développement des produits, notamment en permettant l’optimisation multiphysique en grande dimension. / In this manuscript, we present a shape optimization method based on CAD parameters such as lengths, angles, etc. We rely on gradient-based optimization techniques. The sensitivity of the objective function, with respect to the mesh nodes position, is provided by an adjoint solver considered here as a black box. To optimize with respect to CAD parameters, we focus on computing the sensitivity of the nodes positions with respect to these parameters. Thus, we propose two approaches based on finite differences. The first method uses a harmonic projection to compare in the same space the initial mesh and the one obtained after a change of the set of CAD parameters. The developments presented in this manuscript open up new doors like the application to shapes with multiple borders such as exhaust manifolds. We also developed an interpolation method suitable for this comparison. The entire process is automated, and we demonstrate the entire effectiveness on internal aerodynamics industrial applications. The second method is directly based on the CAD geometries to assess this sensitivity. To perform this comparison, we use the intrinsic definition of the patches in the parametric space (u;v). Through the use of the exact coordinates at any point on the surface provided by the CAD, we avoid using an interpolation to get the best calculation accuracy possible. However, unlike the first method, it requires to identify the correspondence between patches from one shape to another. An application on an external aerodynamics academic case was made. The relevance of the first method is demonstrated on a representative multi-objective case, which facilitate its deployment use in an industrial environment. Regarding the second method, we showed its great potential. However, further developments are needed to handle more advanced cases. Because they are independent of the mechanical solver and the number of parameters, these methods significantly reduce product development time, particularly by allowing large and multiphysics optimization.

No description available.

No description available.

This thesis report aims to explain the research, implementation, and testing of a monocular SLAM system in an application developed by Voysys AB called Oden, as well as the making and investigation of a new data set used to test the SLAM system. Using CUDASIFT to find and match feature points, OpenCV to compute the initial guess, and the Ceres Solver to optimize the results. / <p>Examensarbetet är utfört vid Institutionen för teknik och naturvetenskap (ITN) vid Tekniska fakulteten, Linköpings universitet</p>