Submitted by Rosivalda Pereira (mrs.pereira@ufma.br) on 2017-06-22T19:42:13Z No. of bitstreams: 1 Ezequiel BorgesMelo.pdf: 3711016 bytes, checksum: 4169bfc0d8fb15318ce5eacdd7bda426 (MD5) / Made available in DSpace on 2017-06-22T19:42:13Z (GMT). No. of bitstreams: 1 Ezequiel BorgesMelo.pdf: 3711016 bytes, checksum: 4169bfc0d8fb15318ce5eacdd7bda426 (MD5) Previous issue date: 2015-08-21 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Fundação de Amparo à Pesquisa e ao Desenvolvimento Científico e Tecnológico do Maranhão (FAPEMA) / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPQ) / Amino acids, having both the carboxylic group and amine group, may act as bidentate ligands and, depending on its radical group can also act as tridentate ligands. Amino acids complexed with transition metals have attracted the interest of chemists and physicists because of its possible applications and physical properties. L-threonine complexes with Cu+2 (Cu(II)(L-Threonine)2(H2O)), Co+2 (Co(II)(L-Threonine)2(H2O)2) e Zn+2 (Zn(II)(L-Threonine)2(H2O)2) transition metals already exist in the literature and their crystalline structures are different. However, L-threonine complexed with Ni+2 was not found in the literature. Thus, in this study, we used the amino acid L-threonine as a ligand and Ni+2 ion as the transition metal to obtain the crystal of L-threonine complexed with Ni. For that we used the Slow Evaporation crystal growth method, where a solution containing L-threonine and NiCl2.6H2O with molar ratio (2:1) and NaOH to get a basic pH, is allowed to stand for promoting the crystal growth. To get the crystal structure of this material, X-ray diffraction measures were carried in a APEX2 DUO diffractometer of the Crystallography Laboratory in the Physics Institute of UFG. The data analysis and the resolution of the structure were performed the package Bruker SHELXTL and also using the mechanism of structural determination by Direct Methods, one of the most used ways to overcome the phase problem in the structure determination of small molecules. L-threonine complexed with Ni has the chemical formula Ni(II)(L-Treonina)2(H2O)2 and crystallizes in the orthorhombic system with space group C2221. Thus, we identified that the Ni(II)(L-Treonina)2(H2O)2 crystal has a very similar crystalline structure as Co(II)(L-Treonina)2(H2O)2. Furthermore, the knowledge of the structure of this material opens up a range studies can be performed on it. / Aminoácidos, por terem tanto um grupo carboxílico como um grupo amina, podem agir como ligantes bidentados e, dependendo do seu grupo radical, podem agir também como ligantes tridentados. Aminoácidos complexados com metais de transição têm atraído o interesse de químicos e físicos devido as suas possíveis propriedade físicas e aplicações. Complexos de L-treonina com os metais de transição Cu+2 (Cu(II)(L-Treonina)2(H2O)), Co+2 (Co(II)(L-Treonina)2(H2O)2) e Zn+2 (Zn(II)(L-Treonina)2(H2O)2) já existem na literatura e os três possuem estruturas cristalinas diferentes. Entretanto, a L-treonina complexada com Ni+2 não foi encontrada na literatura. Desta forma, neste trabalho, utilizamos o aminoácido L-treonina como ligante e o íon Ni+2 como metal de transição para obter o cristal de L-treonina complexada com Ni. Para obter estes cristais, utilizamos o método de crescimento por Evaporação Lenta, onde uma solução contendo L-treonina e NiCl2.6H2O na proporção molar (2:1) mais NaOH, para deixar o pH básico, é deixada em repouso para promover o crescimento dos cristais. Para determinar a estrutura cristalina deste material, foram realizados medidas de Difração de Raios X no Difratômetro APEX2 DUO da Bruker, do Laboratório de Cristalografia do Instituto de Física da UFG. O tratamento dos dados e a resolução da estrutura foram realizados usando utilizando o pacote SHELXTL da Bruker e utilizando o mecanismo de determinação estrutural por Métodos Diretos, que é uma das alternativas mais utilizadas para contornar o problema das fases na determinação estrutural de pequenas moléculas. Com essas análises, foi determinado que a L-treonina complexada com Ni possui fórmula química Ni(II)(L-Treonina)2(H2O)2 e cristaliza na simetria ortorrômbica com grupo espacial C2221. Desta forma, identificamos que o cristal de Ni(II)(L-Treonina)2(H2O)2 tem a estrutura cristalina muito similar ao Co(II)(L-Treonina)2(H2O)2. Além disso, com o conhecimento da estrutura desse material, abre-se um leque estudos que podem ser realizados nele.

The problem addressed in this research is the in-flight generation of trajectories for autonomous unmanned aircraft, which requires a method of generating pseudo-optimal trajectories in near-real-time, on-board the aircraft, and without external intervention. The focus of this research is the enhancement of a particular inverse dynamics direct method that is a candidate solution to the problem. This research introduces the following contributions to the method. A quaternion-based inverse dynamics model is introduced that represents all orientations without singularities, permits smooth interpolation of orientations, and generates more accurate controls than the previous Euler-angle model. Algorithmic modifications are introduced that: overcome singularities arising from parameterization and discretization; combine analytic and finite difference expressions to improve the accuracy of controls and constraints; remove roll ill-conditioning when the normal load factor is near zero, and extend the method to handle negative-g orientations. It is also shown in this research that quadratic interpolation improves the accuracy and speed of constraint evaluation. The method is known to lead to a multimodal constrained nonlinear optimization problem. The performance of the method with four nonlinear programming algorithms was investigated: a differential evolution algorithm was found to be capable of over 99% successful convergence, to generate solutions with better optimality than the quasi- Newton and derivative-free algorithms against which it was tested, but to be up to an order of magnitude slower than those algorithms. The effects of the degree and form of polynomial airspeed parameterization on optimization performance were investigated, and results were obtained that quantify the achievable optimality as a function of the parameterization degree. Overall, it was found that the method is a potentially viable method of on-board near- real-time trajectory generation for unmanned aircraft but for this potential to be realized in practice further improvements in computational speed are desirable. Candidate optimization strategies are identified for future research.

Este trabalho está dividido em duas partes distintas que constituem contribuições inéditas ao estudo da estabilidade em sistemas elétricos de potência. A primeira parte do trabalho é a mais importante e trata do problema da identificação das soluções de baixa tensão críticas do fluxo de carga. Esta parte do trabalho se presta a análise de estabilidade de tensão a pequenas perturbações. Os últimos capítulos deste trabalho apresentam também uma proposta de função energia estendida que modela as cargas dependentes da tensão segundo o modelo ZIP de carga, considerando a estrutura da rede preservada. Assim, a função energia proposta pode ser utilizada para analisar tanto a estabilidade de tensão como a estabilidade de ângulo em sistemas de potência. Esta proposta também é inédita na literatura. Embora a função energia proposta tenha sido aplicada apenas a sistemas de dimensão reduzidas, os resultados apresentados neste trabalho nos levam a acreditar que essa mesma função energia pode ser utilizada na análise de estabilidade de sistemas de potência de grandes dimensões. Já o método proposto para identificação das soluções de baixa tensão das equações de fluxo de carga se utiliza de um sistema dinâmico auxiliar das equações de fluxo de carga. O sistema dinâmico auxiliar utilizado não tem significado físico, mas pode ser escolhido de tal forma que a solução usual das equações de fluxo de carga seja um ponto de equilíbrio estável do sistema dinâmico auxiliar, eque as soluções de baixa tensão do fluxo de carga sejam pontos de equilíbrio instáveis do sistema dinâmico auxiliar. Dessa forma, é possível calcular as soluções de baixa tensão do fluxo de carga, calculando-se os pontos de equilíbrio instáveis do sistema dinâmico auxiliar. Assim, é possível utilizar partes da teoria de sistemas dinâmicos para estudar as soluções das equações de fluxo de carga. Baseado nestes princípios, foi desenvolvido um programa para calcular trajetórias do sistema dinâmico auxiliar, que se iniciam e se mantêm nas vizinhanças da fronteira da área de atração do ponto de equilíbrio estável do SEP. Dessa forma é possível afirmar que a trajetória calculada tende a convergir para a solução crítica das equações de fluxo de carga. O programa foi inicialmente concebido para calcular as soluções de baixa tensão de sistemas elétricos sem perdas. Em seguida o programa desenvolvido foi adaptado para calcular as soluções de baixa tensão de sistemas de potência completos, incluindo também as resistências das linhas de transmissão. Esta última versão do programa foi testada para os sistemas IEEE 39 e IEEE 118 barras, e os resultados obtidos se mostraram bastante satisfatórios. Assim, o método proposto é uma ferramenta original e eficaz para a solução do problema de calcular a solução crítica das equações de fluxo de carga de sistemas elétricos de potência. / This work may be divided into two distinct parts. Both of them are new contributions to stability analysis of power systems. In the first part it is proposed a new method to calculate the critical load flow low voltage solutions, and it is the main part of this work. Meanwhile, the last two chapters of this work presents a proposed extended energy function that consider the common load ZIP models. It allows the analysis of angle and voltage stability for power systems subjected to large disturbances. This work proposes a method to calculate the low voltage solutions (LVS) of the load flow equations of an electrical power system. The proposed method identifies the LVS involved in the saddle-node bifurcation leading the power system to a voltage collapse. This solution is known as the critical low voltage solution. In order to perform the proposed calculation, an auxiliary dynamical gradient system is used. It is shown that the equilibrium points of that associated auxiliary dynamical gradient system are the solutions of the load flow equations. In such manner, the paper proposes identifying the critical LVS calculating the equilibrium points of an auxiliary dynamical gradient system. The proposed method was tested on the Stagg 5-bus, on the IEEE 39-bus and on IEEE 118-bus test systems, and the results are presented at the end of the text.

Optimal Control (OC) is essentially a mathematical extremal problem. The procedure consists on the definition of a criterion to minimize (or maximize), some constraints that must be fulfilled and boundary conditions or disturbances affecting to the system behavior. The OC theory supplies methods to derive a control trajectory that minimizes (or maximizes) that criterion. This dissertation addresses the application of OC to automotive control problems at the powertrain level, with emphasis on the internal combustion engine. The necessary tools are an optimization method and a mathematical representation of the powertrain. Thus, the OC theory is reviewed with a quantitative analysis of the advantages and drawbacks of the three optimization methods available in literature: dynamic programming, Pontryagin minimum principle and direct methods. Implementation algorithms for these three methods are developed and described in detail. In addition to that, an experimentally validated dynamic powertrain model is developed, comprising longitudinal vehicle dynamics, electrical motor and battery models, and a mean value engine model. OC can be utilized for three different purposes: 1. Applied control, when all boundaries can be accurately defined. The engine control is addressed with this approach assuming that a the driving cycle is known in advance, translating into a large mathematical problem. Two specific cases are studied: the management of a dual-loop EGR system, and the full control of engine actuators, namely fueling rate, SOI, EGR and VGT settings. 2. Derivation of near-optimal control rules, to be used if some disturbances are unknown. In this context, cycle-specific engine calibrations calculation, and a stochastic feedback control for power-split management in hybrid vehicles are analyzed. 3. Use of OC trajectories as a benchmark or base line to improve the system design and efficiency with an objective criterion. OC is used to optimize the heat release law of a diesel engine and to size a hybrid powertrain with a further cost analysis. OC strategies have been applied experimentally in the works related to the internal combustion engine, showing significant improvements but non-negligible difficulties, which are analyzed and discussed. The methods developed in this dissertation are general and can be extended to other criteria if appropriate models are available. / El Control Óptimo (CO) es esencialmente un problema matemático de búsqueda de extremos, consistente en la definición de un criterio a minimizar (o maximizar), restricciones que deben satisfacerse y condiciones de contorno que afectan al sistema. La teoría de CO ofrece métodos para derivar una trayectoria de control que minimiza (o maximiza) ese criterio. Esta Tesis trata la aplicación del CO en automoción, y especialmente en el motor de combustión interna. Las herramientas necesarias son un método de optimización y una representación matemática de la planta motriz. Para ello, se realiza un análisis cuantitativo de las ventajas e inconvenientes de los tres métodos de optimización existentes en la literatura: programación dinámica, principio mínimo de Pontryagin y métodos directos. Se desarrollan y describen los algoritmos para implementar estos métodos así como un modelo de planta motriz, validado experimentalmente, que incluye la dinámica longitudinal del vehículo, modelos para el motor eléctrico y las baterías, y un modelo de motor de combustión de valores medios. El CO puede utilizarse para tres objetivos distintos: 1. Control aplicado, en caso de que las condiciones de contorno estén definidas. Puede aplicarse al control del motor de combustión para un ciclo de conducción dado, traduciéndose en un problema matemático de grandes dimensiones. Se estudian dos casos particulares: la gestión de un sistema de EGR de doble lazo, y el control completo del motor, en particular de las consignas de inyección, SOI, EGR y VGT. 2. Obtención de reglas de control cuasi-óptimas, aplicables en casos en los que no todas las perturbaciones se conocen. A este respecto, se analizan el cálculo de calibraciones de motor específicas para un ciclo, y la gestión energética de un vehículo híbrido mediante un control estocástico en bucle cerrado. 3. Empleo de trayectorias de CO como comparativa o referencia para tareas de diseño y mejora, ofreciendo un criterio objetivo. La ley de combustión así como el dimensionado de una planta motriz híbrida se optimizan mediante el uso de CO. Las estrategias de CO han sido aplicadas experimentalmente en los trabajos referentes al motor de combustión, poniendo de manifiesto sus ventajas sustanciales, pero también analizando dificultades y líneas de actuación para superarlas. Los métodos desarrollados en esta Tesis Doctoral son generales y aplicables a otros criterios si se dispone de los modelos adecuados. / El Control Òptim (CO) és essencialment un problema matemàtic de cerca d'extrems, que consisteix en la definició d'un criteri a minimitzar (o maximitzar), restriccions que es deuen satisfer i condicions de contorn que afecten el sistema. La teoria de CO ofereix mètodes per a derivar una trajectòria de control que minimitza (o maximitza) aquest criteri. Aquesta Tesi tracta l'aplicació del CO en automoció i especialment al motor de combustió interna. Les ferramentes necessàries són un mètode d'optimització i una representació matemàtica de la planta motriu. Per a això, es realitza una anàlisi quantitatiu dels avantatges i inconvenients dels tres mètodes d'optimització existents a la literatura: programació dinàmica, principi mínim de Pontryagin i mètodes directes. Es desenvolupen i descriuen els algoritmes per a implementar aquests mètodes així com un model de planta motriu, validat experimentalment, que inclou la dinàmica longitudinal del vehicle, models per al motor elèctric i les bateries, i un model de motor de combustió de valors mitjans. El CO es pot utilitzar per a tres objectius diferents: 1. Control aplicat, en cas que les condicions de contorn estiguen definides. Es pot aplicar al control del motor de combustió per a un cicle de conducció particular, traduint-se en un problema matemàtic de grans dimensions. S'estudien dos casos particulars: la gestió d'un sistema d'EGR de doble llaç, i el control complet del motor, particularment de les consignes d'injecció, SOI, EGR i VGT. 2. Obtenció de regles de control quasi-òptimes, aplicables als casos on no totes les pertorbacions són conegudes. A aquest respecte, s'analitzen el càlcul de calibratges específics de motor per a un cicle, i la gestió energètica d'un vehicle híbrid mitjançant un control estocàstic en bucle tancat. 3. Utilització de trajectòries de CO com comparativa o referència per a tasques de disseny i millora, oferint un criteri objectiu. La llei de combustió així com el dimensionament d'una planta motriu híbrida s'optimitzen mitjançant l'ús de CO. Les estratègies de CO han sigut aplicades experimentalment als treballs referents al motor de combustió, manifestant els seus substancials avantatges, però també analitzant dificultats i línies d'actuació per superar-les. Els mètodes desenvolupats a aquesta Tesi Doctoral són generals i aplicables a uns altres criteris si es disposen dels models adequats. / Reig Bernad, A. (2017). Optimal Control for Automotive Powertrain Applications [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/90624 / TESIS

Nous nous intéressons à la résolution de systèmes linéaires creux de très grande taille sur des machines parallèles. Dans ce contexte, la mémoire est un facteur qui limite voire empêche souvent l’utilisation de solveurs directs, notamment ceux basés sur la méthode multifrontale. Cette étude se concentre sur les problèmes de mémoire et de performance des deux phases des méthodes directes les plus coûteuses en mémoire et en temps : la factorisation numérique et la résolution triangulaire. Dans une première partie nous nous intéressons à la phase de résolution à seconds membres creux, puis, dans une seconde partie, nous nous intéressons à la scalabilité mémoire de la factorisation multifrontale. La première partie de cette étude se concentre sur la résolution triangulaire à seconds membres creux, qui apparaissent dans de nombreuses applications. En particulier, nous nous intéressons au calcul d’entrées de l’inverse d’une matrice creuse, où les seconds membres et les vecteurs solutions sont tous deux creux. Nous présentons d’abord plusieurs schémas de stockage qui permettent de réduire significativement l’espace mémoire utilisé lors de la résolution, dans le cadre d’exécutions séquentielles et parallèles. Nous montrons ensuite que la façon dont les seconds membres sont regroupés peut fortement influencer la performance et nous considérons deux cadres différents : le cas "hors-mémoire" (out-of-core) où le but est de réduire le nombre d’accès aux facteurs, qui sont stockés sur disque, et le cas "en mémoire" (in-core) où le but est de réduire le nombre d’opérations. Finalement, nous montrons comment améliorer le parallélisme. Dans la seconde partie, nous nous intéressons à la factorisation multifrontale parallèle. Nous montrons tout d’abord que contrôler la mémoire active spécifique à la méthode multifrontale est crucial, et que les technique de "répartition" (mapping) classiques ne peuvent fournir une bonne scalabilité mémoire : le coût mémoire de la factorisation augmente fortement avec le nombre de processeurs. Nous proposons une classe d’algorithmes de répartition et d’ordonnancement "conscients de la mémoire" (memory-aware) qui cherchent à maximiser la performance tout en respectant une contrainte mémoire fournie par l’utilisateur. Ces techniques ont révélé des problèmes de performances dans certains des noyaux parallèles denses utilisés à chaque étape de la factorisation, et nous avons proposé plusieurs améliorations algorithmiques. Les idées présentées tout au long de cette étude ont été implantées dans le solveur MUMPS (Solveur MUltifrontal Massivement Parallèle) et expérimentées sur des matrices de grande taille (plusieurs dizaines de millions d’inconnues) et sur des machines massivement parallèles (jusqu’à quelques milliers de coeurs). Elles ont permis d’améliorer les performances et la robustesse du code et seront disponibles dans une prochaine version. Certaines des idées présentées dans la première partie ont également été implantées dans le solveur PDSLin (solveur linéaire hybride basé sur une méthode de complément de Schur). / We consider the solution of very large sparse systems of linear equations on parallel architectures. In this context, memory is often a bottleneck that prevents or limits the use of direct solvers, especially those based on the multifrontal method. This work focuses on memory and performance issues of the two memory and computationally intensive phases of direct methods, that is, the numerical factorization and the solution phase. In the first part we consider the solution phase with sparse right-hand sides, and in the second part we consider the memory scalability of the multifrontal factorization. In the first part, we focus on the triangular solution phase with multiple sparse right-hand sides, that appear in numerous applications. We especially emphasize the computation of entries of the inverse, where both the right-hand sides and the solution are sparse. We first present several storage schemes that enable a significant compression of the solution space, both in a sequential and a parallel context. We then show that the way the right-hand sides are partitioned into blocks strongly influences the performance and we consider two different settings: the out-of-core case, where the aim is to reduce the number of accesses to the factors, that are stored on disk, and the in-core case, where the aim is to reduce the computational cost. Finally, we show how to enhance the parallel efficiency. In the second part, we consider the parallel multifrontal factorization. We show that controlling the active memory specific to the multifrontal method is critical, and that commonly used mapping techniques usually fail to do so: they cannot achieve a high memory scalability, i.e. they dramatically increase the amount of memory needed by the factorization when the number of processors increases. We propose a class of "memory-aware" mapping and scheduling algorithms that aim at maximizing performance while enforcing a user-given memory constraint and provide robust memory estimates before the factorization. These techniques have raised performance issues in the parallel dense kernels used at each step of the factorization, and we have proposed some algorithmic improvements. The ideas presented throughout this study have been implemented within the MUMPS (MUltifrontal Massively Parallel Solver) solver and experimented on large matrices (up to a few tens of millions unknowns) and massively parallel architectures (up to a few thousand cores). They have demonstrated to improve the performance and the robustness of the code, and will be available in a future release. Some of the ideas presented in the first part have also been implemented within the PDSLin (Parallel Domain decomposition Schur complement based Linear solver) solver.

O presente trabalho tem como objetivo investigar a presença de diferentes escalas de tempo nos modelos matemáticos que descrevem a dinâmica dos sistemas elétricos de potência (SEPs), em particular a existência de duas escalas de tempo distintas: lenta e rápida, e explorá-las no estudo de estabilidade transitória destes sistemas através da utilização de métodos diretos (funções energia). Em particular, o método do Ponto de Equilíbrio Instável de Controle (CUEP) para modelos com duas escalas de tempo será estudado e aplicado na análise de estabilidade transitória de SEPs. As bases teóricas para a análise de estabilidade transitória, de sistemas com duas escalas de tempo, serão apresentadas, assim como funções energia e novos algoritmos numéricos para o cálculo do CUEP nestes sistemas, a fim de evidenciar as melhorias e possíveis limitações deste novo método CUEP em duas escalas de tempo quando comparado ao método CUEP tradicional. Explorando as escalas de tempo lenta e rápida na análise de estabilidade transitória, espera-se que novos algoritmos numéricos mais robustos para o cálculo do CUEP sejam obtidos, assim como a diminuição do conservadorismo dos resultados. / The main objective of this work is to investigate the existence of several time-scales in the mathematical models of electric power systems, in particular the existence of two-time scales: slow and fast, and exploit these features in the direct transient stability assessment. In particular, the Controlling Unstable Equilibrium Point (CUEP) method is studied for two-time scale models of power systems and applied to transient stability analysis. In order to accomplish this aim, a sound theoretical basis for two-time scale transient stability analysis of electric power system models will be provided, as well as energy functions and new numerical algorithms for proper two-time scale CUEP calculations, with the purpose of investigating improvements and possible limitations of this method when compared with the traditional CUEP method. Exploiting the two-time scale features of power system models, it is intended to obtain new robust numerical algorithms for transient stability analysis, as well as to diminish the conservativeness of the results.

O presente trabalho tem como objetivo investigar a presença de diferentes escalas de tempo nos modelos matemáticos que descrevem a dinâmica dos sistemas elétricos de potência (SEPs), em particular a existência de duas escalas de tempo distintas: lenta e rápida, e explorá-las no estudo de estabilidade transitória destes sistemas através da utilização de métodos diretos (funções energia). Em particular, o método do Ponto de Equilíbrio Instável de Controle (CUEP) para modelos com duas escalas de tempo será estudado e aplicado na análise de estabilidade transitória de SEPs. As bases teóricas para a análise de estabilidade transitória, de sistemas com duas escalas de tempo, serão apresentadas, assim como funções energia e novos algoritmos numéricos para o cálculo do CUEP nestes sistemas, a fim de evidenciar as melhorias e possíveis limitações deste novo método CUEP em duas escalas de tempo quando comparado ao método CUEP tradicional. Explorando as escalas de tempo lenta e rápida na análise de estabilidade transitória, espera-se que novos algoritmos numéricos mais robustos para o cálculo do CUEP sejam obtidos, assim como a diminuição do conservadorismo dos resultados. / The main objective of this work is to investigate the existence of several time-scales in the mathematical models of electric power systems, in particular the existence of two-time scales: slow and fast, and exploit these features in the direct transient stability assessment. In particular, the Controlling Unstable Equilibrium Point (CUEP) method is studied for two-time scale models of power systems and applied to transient stability analysis. In order to accomplish this aim, a sound theoretical basis for two-time scale transient stability analysis of electric power system models will be provided, as well as energy functions and new numerical algorithms for proper two-time scale CUEP calculations, with the purpose of investigating improvements and possible limitations of this method when compared with the traditional CUEP method. Exploiting the two-time scale features of power system models, it is intended to obtain new robust numerical algorithms for transient stability analysis, as well as to diminish the conservativeness of the results.

Nous considérons la résolution de très grands systèmes linéaires creux à l'aide d'une méthode de factorisation directe appelée méthode multifrontale. Bien que numériquement robustes et faciles à utiliser (elles ne nécessitent que des informations algébriques : la matrice d'entrée A et le second membre b, même si elles peuvent exploiter des stratégies de prétraitement basées sur des informations géométriques), les méthodes directes sont très coûteuses en termes de mémoire et d'opérations, ce qui limite leur applicabilité à des problèmes de taille raisonnable (quelques millions d'équations). Cette étude se concentre sur l'exploitation des approximations de rang-faible dans la méthode multifrontale, pour réduire sa consommation mémoire et son volume d'opérations, dans des environnements séquentiel et à mémoire distribuée, sur une large classe de problèmes. D'abord, nous examinons les formats rang-faible qui ont déjà été développé pour représenter efficacement les matrices denses et qui ont été utilisées pour concevoir des solveurs rapides pour les équations aux dérivées partielles, les équations intégrales et les problèmes aux valeurs propres. Ces formats sont hiérarchiques (les formats H et HSS sont les plus répandus) et il a été prouvé, en théorie et en pratique, qu'ils permettent de réduire substantiellement les besoins en mémoire et opération des calculs d'algèbre linéaire. Cependant, de nombreuses contraintes structurelles sont imposées sur les problèmes visés, ce qui peut limiter leur efficacité et leur applicabilité aux solveurs multifrontaux généraux. Nous proposons un format plat appelé Block Rang-Faible (BRF) basé sur un découpage naturel de la matrice en blocs et expliquons pourquoi il fournit toute la flexibilité nécéssaire à son utilisation dans un solveur multifrontal général, en terme de pivotage numérique et de parallélisme. Nous comparons le format BRF avec les autres et montrons que le format BRF ne compromet que peu les améliorations en mémoire et opération obtenues grâce aux approximations rang-faible. Une étude de stabilité montre que les approximations sont bien contrôlées par un paramètre numérique explicite appelé le seuil rang-faible, ce qui est critique dans l'optique de résoudre des systèmes linéaires creux avec précision. Ensuite, nous expliquons comment les factorisations exploitant le format BRF peuvent être efficacement implémentées dans les solveurs multifrontaux. Nous proposons plusieurs algorithmes de factorisation BRF, ce qui permet d'atteindre différents objectifs. Les algorithmes proposés ont été implémentés dans le solveur multifrontal MUMPS. Nous présentons tout d'abord des expériences effectuées avec des équations aux dérivées partielles standardes pour analyser les principales propriétés des algorithmes BRF et montrer le potentiel et la flexibilité de l'approche ; une comparaison avec un code basé sur le format HSS est également fournie. Ensuite, nous expérimentons le format BRF sur des problèmes variés et de grande taille (jusqu'à une centaine de millions d'inconnues), provenant de nombreuses applications industrielles. Pour finir, nous illustrons l'utilisation de notre approche en tant que préconditionneur pour la méthode du Gradient Conjugué. / We consider the solution of large sparse linear systems by means of direct factorization based on a multifrontal approach. Although numerically robust and easy to use (it only needs algebraic information: the input matrix A and a right-hand side b, even if it can also digest preprocessing strategies based on geometric information), direct factorization methods are computationally intensive both in terms of memory and operations, which limits their scope on very large problems (matrices with up to few hundred millions of equations). This work focuses on exploiting low-rank approximations on multifrontal based direct methods to reduce both the memory footprints and the operation count, in sequential and distributed-memory environments, on a wide class of problems. We first survey the low-rank formats which have been previously developed to efficiently represent dense matrices and have been widely used to design fast solutions of partial differential equations, integral equations and eigenvalue problems. These formats are hierarchical (H and Hierarchically Semiseparable matrices are the most common ones) and have been (both theoretically and practically) shown to substantially decrease the memory and operation requirements for linear algebra computations. However, they impose many structural constraints which can limit their scope and efficiency, especially in the context of general purpose multifrontal solvers. We propose a flat format called Block Low-Rank (BLR) based on a natural blocking of the matrices and explain why it provides all the flexibility needed by a general purpose multifrontal solver in terms of numerical pivoting for stability and parallelism. We compare BLR format with other formats and show that BLR does not compromise much the memory and operation improvements achieved through low-rank approximations. A stability study shows that the approximations are well controlled by an explicit numerical parameter called low-rank threshold, which is critical in order to solve the sparse linear system accurately. Details on how Block Low-Rank factorizations can be efficiently implemented within multifrontal solvers are then given. We propose several Block Low-Rank factorization algorithms which allow for different types of gains. The proposed algorithms have been implemented within the MUMPS (MUltifrontal Massively Parallel Solver) solver. We first report experiments on standard partial differential equations based problems to analyse the main features of our BLR algorithms and to show the potential and flexibility of the approach; a comparison with a Hierarchically SemiSeparable code is also given. Then, Block Low-Rank formats are experimented on large (up to a hundred millions of unknowns) and various problems coming from several industrial applications. We finally illustrate the use of our approach as a preconditioning method for the Conjugate Gradient.

L'objectif premier de cette thèse est d'étudier certains aspects du contrôle d'attitude d'un corps rigide, afin d'optimiser la trajectoire d'un lanceur au cours de sa phase balistique. Nous y développons un cadre mathématique permettant de formuler ce problème comme un problème de contrôle optimal avec des contraintes intermédiaires sur l'état. En parallèle de l'étude théorique de ce problème, nous avons mené l'implémentation d'un logiciel d'optimisation basé sur la combinaison d'une méthode directe et d'un algorithme de point intérieur, permettant à l'utilisateur de traiter une phase balistique quelconque. Nous entendons par là qu'il est possible de spécifier un nombre quelconque de contraintes intermédiaires, correspondant à un nombre quelconque de largages de charges utiles. En outre, nous avons appliqué les méthodes dites indirectes, exploitant le principe du maximum de Pontryagin, à la résolution de ce problème de contrôle optimal. On cherche dans ce travail à trouver des trajectoires optimales du point de vue de la consommation en ergols, ce qui correspond à un coût L 1 . Réputé difficile numériquement, ce critère peut être atteint grâce à une méthode de continuation, en se servant d'un coût L 2 comme intermédiaire de calcul et en déformant progressivement ce problème L 2 . Nous verrons également d'autres exemples d'application des méthodes de continuation. Enfin, nous présenterons également un algorithme de contrôle robuste, permettant de rejoindre un état cible à partir d'un état perturbé, en suivant une trajectoire de référence tout en conservant la structure bang-bang des contrôles. La robustesse d'un contrôle peut également être améliorée par l'ajout de variations aiguilles, et un critère qualifiant la robustesse d'une trajectoire à partir des valeurs singulières d'une certaine application entrée-sortie est déduit. / The first objective of this work is to study some aspects of the attitude control problem of a rigid body, in order to optimize the trajectory of a launcher during a ballistic flight. We state this problem in a general mathematical setting, as an optimal control problem with intermediate constraints on the state. Meanwhile, we also implement an optimization software that relies on the combination of a direct method and of an interior-point algorithm to optimize any given ballistic flight, with any number of intermediate constraints, corresponding to any number of satellite separations. Besides, we applied the so-called indirect methods, exploiting Pontryagin maximum principle, to the resolution of this optimal control problem. In this work, optimal trajectories with respect to the consumption are looked after, which corresponds to a L 1 cost. Known to be numerically challenging, this criterion can be reached by performing a continuation procedure, starting from a L 2 cost, for which it is easier to provide a good initialization of the underlying optimization algorithm. We shall also study other examples of applications for continuation procedures. Eventually, we will present a robust control algorithm, allowing to reach a target point from a perturbed initial point, following a nominal trajectory while preserving its bang-bang structure. The robustness of a control can be improved introducing needle-like variations, and a criterion to measure the robustness of a trajectory is designed, involving the singular value decomposition of some end-point mapping.

<div>Optimal control theory (OCT) exists since the 1950s. However, with the advent of modern computers, the design community delegated the task of solving the optimal control problems (OCPs) largely to computationally intensive direct methods instead of methods that use OCT. Some recent work showed that solvers using OCT could leverage parallel computing resources for faster execution. The need for near real-time, high quality solutions for OCPs has therefore renewed interest in OCT in the design community. However, certain challenges still exist that prohibits its use for solving complex practical aerospace problems, such as landing human-class payloads safely on Mars.</div><div><br></div><div>In order to advance OCT, this thesis introduces Epsilon-Trig regularization method to simply and efficiently solve bang-bang and singular control problems. The Epsilon-Trig method resolves the issues pertaining to the traditional smoothing regularization method. Some benchmark problems from the literature including the Van Der Pol oscillator, the boat problem, and the Goddard rocket problem verified and validated the Epsilon-Trig regularization method using GPOPS-II.</div><div><br></div><div>This study also presents and develops the usage of trigonometry for incorporating control bounds and mixed state-control constraints into OCPs and terms it as Trigonometrization. Results from literature and GPOPS-II verified and validated the Trigonometrization technique using certain benchmark OCPs. Unlike traditional OCT, Trigonometrization converts the constrained OCP into a two-point boundary value problem rather than a multi-point boundary value problem, significantly reducing the computational effort required to formulate and solve it. This work uses Trigonometrization to solve some complex aerospace problems including prompt global strike, noise-minimization for general aviation, shuttle re-entry problem, and the g-load constraint problem for an impactor. Future work for this thesis includes the development of the Trigonometrization technique for OCPs with pure state constraints.</div>