A designação de estruturas próprias de sistemas dinâmicos com retroação completa de estados ou saídas foi objeto de estudo de muitos pesquisadores durante a segunda metade do século XX. Os trabalhos mais relevantes sobre o tema foram revisados e serviram como base para a elaboração da metodologia apresentada neste trabalho. Que consiste na designação de estruturas próprias desejadas para um sistema linear em malha aberta com a modificação de parâmetros do sistema através da pseudo retroação de saídas devidamente definidas. O método foi aplicado na coxinização de um veículo de passeio. No qual os modos de vibrar de modelos lineares com 7 e 8 graus de liberdade do veículo foram adequados com o intuito de reduzir as acelerações verticais de chassi, características do fenômeno shake de motor e câmbio (faixa de frequência entre 7 e 25 Hz). Para tanto, reduziu-se a participação do grau de liberdade vertical de chassi nos modos com grande participação dos graus de liberdade de motor e massa não suspensa. Os resultados obtidos com a aplicação da metodologia na coxinização foram valores de rigidezes, amortecimentos e características de coxins hidráulicos que resultam na redução significativa da aceleração vertical de chassi, que implica em uma melhora perceptível para o consumidor na qualidade do conforto do veículo. / Eigenstructure assignment using full state or output feedback control had been object of study of many researchers during the second half of XX century. The most important works about eigenstructure assignment were reviewed, specially some applications within aerospace industry, that was the responsible for developing all the theory concerning pole and eigenvector placement. The review of the related theory was also based on the pioneering and most expressive works and were base for the methodology developed and described in this work. Which consists basically on the assignment of some desired eigensctructure of open loop linear systems modifying some parameters of the systems by means pseudo feedback of some outputs into inputs specifically defined. This methodology was applied to a 8 DOF vehicle model, a case of study, in order to adequate the system modes changing engine mounts parameters to improve the vehicle ride comfort, mainly eigenstructures about powertrain shake range frequencies.

Les nouvelles architectures de calcul intensif intègrent de plus en plus de microprocesseurs qui eux-mêmes intègrent un nombre croissant de cœurs de calcul. Cette multiplication des unités de calcul dans les architectures ont fait apparaître des topologies fortement hiérarchiques. Ces architectures sont dites NUMA. Les algorithmes de simulation numérique et les solveurs de systèmes linéaires qui en sont une brique de base doivent s'adapter à ces nouvelles architectures dont les accès mémoire sont dissymétriques. Nous proposons dans cette thèse d'introduire un ordonnancement dynamique adapté aux architectures NUMA dans le solveur PaStiX. Les structures de données du solveur, ainsi que les schémas de communication ont dû être modifiés pour répondre aux besoins de ces architectures et de l'ordonnancement dynamique. Nous nous sommes également intéressés à l'adaptation dynamique du grain de calcul pour exploiter au mieux les architectures multi-cœurs et la mémoire partagée. Ces développements sont ensuite validés sur un ensemble de cas tests sur différentes architectures. / New supercomputers incorporate many microprocessors which include themselves one or many computational cores. These new architectures induce strongly hierarchical topologies. These are called NUMA architectures. Sparse direct solvers are a basic building block of many numerical simulation algorithms. They need to be adapted to these new architectures with Non Uniform Memory Accesses. We propose to introduce a dynamic scheduling designed for NUMA architectures in the PaStiX solver. The data structures of the solver, as well as the patterns of communication have been modified to meet the needs of these architectures and dynamic scheduling. We are also interested in the dynamic adaptation of the computation grain to use efficiently multi-core architectures and shared memory. Experiments on several numerical test cases will be presented to prove the efficiency of the approach on different architectures.

En esta tesis, se utilizan t ecnicas matem atico-epidemiol ogicas para modelar el rendimiento acad emico en Espa~na (prestando especial atenci on en el fracaso escolar) para comprender mejor los mecanismos detr as de esta importante cuesti on, as como para predecir c omo evolucionar an los resultados acad emicos en el Bachillerato espa~nol en los pr oximos a~nos. El nivel educativo de Bachillerato en Espa~na est a formado por los dos ultimos cursos antes de acceder a la universidad o al mercado de trabajo y corresponde a los estudiantes de 16 18 a~nos. Este nivel educativo es muy importante para la formaci on de los estudiantes ya que representa un periodo en el que deber an tomar importantes decisiones sobre el futuro acad emico y profesional. En primer lugar, en el Cap tulo 2, se presenta un modelo determinista donde se analiza el rendimiento acad emico asumiendo que la actitud negativa de los alumnos de Bachillerato puede ser debida a su comportamiento aut onomo y la in uencia de compa~neros con malos resultados acad emicos. Luego, en el Cap tulo 3, se mejora el modelo basado en la idea de que no s olo los malos h abitos acad emicos se transmiten socialmente sino tambi en los buenos h abitos de estudio. Adem as, descomponemos los par ametros de transmisi on de h abitos acad emicos con el n de analizar con m as detalle qu e grupos de estudiantes son m as susceptibles a ser in uenciados por compa~neros con buenos o malos h abitos acad emicos. El abandono escolar tambi en han sido incluido en este modelo. El enfoque adoptado permite proporcionar predicciones deterministas y con intervalos de con anza de la evoluci on del rendimiento escolar (incluyendo las tasas de abandono) en Bachillerato en Espa~na en los pr oximos a~nos. Este enfoque, adem as, nos permite modelar el rendimiento acad emico en otros niveles educativos del sistema acad emico espa~nol o de fuera de Espa~na tal y como se muestra en el Cap tulo 4, donde el modelo se aplica satisfactoriamente al sistema acad emico actual de la regi on alemana de Renania del Norte-Westfalia. Para concluir esta tesis, proporcionamos una estimaci on de los costes relacionados con el rendimiento acad emico espa~nol basado en nuestras predicciones. Esta estimaci on representa la inversi on en Bachillerato por parte del Gobierno espa~nol y las familias en los pr oximos a~nos, con especial atenci on en los grupos de estudiantes que no promocionan y abandonan en los diferentes cursos acad emicos. / In this dissertation, we use epidemiologic-mathematical techniques to model the academic performance in Spain (paying special attention on the academic underachievement) to understand better the mechanisms behind this important issue as well as to predict how academic results will evolve in the Spanish Bachillerato over the next few years. The Spanish Bachillerato educational level is made up of the last courses before accessing to the university or to the work market and corresponds to students of 16¿18 years old. This educational level is a milestone in the career training of students because it represents a period to make important decisions about academic and professional future. In a rst step, in the Chapter 2 we will present a deterministic model where academic performance is analyzed assuming the negative attitude of Bachillerato students may be due to their autonomous behavior and the in uence of classmates with bad academic results. Then, in the Chapter 3, the model is improved based on the idea that not only the bad academic habits are socially transmitted but also the good study habits. Besides, we decompose the transmission academic habits into good and bad academic habits, in order to analyze with more detail which group of students are more susceptible to be in uenced by good or bad academic students. The consideration of quantifying the abandon rates is also a new issue dealt with in it. The adopted approach allow to provide both punctual and con dence intervals predictions to the evolution of academic performance (including the abandon rates) in Bachillerato in Spain over the next few years. The adopted approach allows us to model academic performance in academic levels other than Bachillerato and/or beyond the Spanish academic system. This issue is assessed in Chapter 4, where the model is satisfactorily applied to the current academic system of the German region of North Rhine-Westphalia. To conclude this dissertation, we provide an estimation of the cost related to the Spanish academic underachievement based on our predictions. This estimation represents the investment in the Spanish Bachillerato from the Spanish Government and families over the next few years, paying special attention on the groups of students who do not promote and abandon during their corresponding academic year. / Sánchez Sánchez, A. (2013). Modelling the evolution dynamics of the academic performance in high school in Spain. Probabilistic predictions of future trends and their economical consequences [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/32280 / TESIS

Les simulateurs d'accouchement sont des outils dédiés à la formation du personnel médical en gynécologie obstétrique. Le BirthSIM développé au laboratoire Ampère constitue une interface haptique de simulateur d'accouchement permettant de reproduire divers scénarios d'accouchement et de former les jeunes sages-femmes et obstétriciens aux différents gestes techniques nécessaires dans le cadre de leurs professions. Dans ce manuscrit, les récentes améliorations apportées au prototype du BirthSIM seront présentées. Parmi ces améliorations, deux sont particulièrement remarquables. La première concerne l'augmentation du nombre de degrés de liberté pilotés de l'interface. Ainsi, les travaux ayant permis le passage d'une interface ne comptant qu'un seul degré de liberté à quatre seront détaillés. Le BirthSIM est le seul simulateur d'accouchement qui intègre à la fois des actionneurs électriques et pneumatiques. La seconde évolution concerne le développement de lois de commande permettant d'améliorer le rendu haptique de l'interface. Les lois de commande retenues permettent d'asservir en position le BirthSIM, mais également de régler la raideur de l'interface. Ce réglage de raideur en coordonnées cartésiennes a pour objectif de reproduire le comportement de la tête fœtale en contact avec les tissus mous du bassin au cours d'une contraction. La loi de commande présentée dans ces travaux de thèse est une loi de commande non-linéaire synthétisée par backstepping. En effet, cette méthode permet de prendre en compte les phénomènes non-linéaires agissant sur le comportement des actionneurs pneumatiques. Cette commande, assurant le suivi d'une consigne de position et une modification de la raideur en temps réel à partir d'une stratégie de réglage de gain, a été testée à la fois en simulation et sur le prototype du BirthSIM. Les performances de cette commande sont également comparées à celles de lois de commande plus classiquement utilisées en robotique. / The childbirth simulators are developed for practitioner training in gynecology and obstetrics. The BirthSIM is a haptic interface which simulates various childbirth scenarios to teach obstetricians and midwives the technical delivery gestures. In this work, the latest improvements on the BirthSIM prototype will be presented. Two improvements are especially detailed. The first one consists in adding actuated degrees of freedom to the haptic interface. Thus, the works which have led to a new four degrees of freedom robot are presented. The BirthSIM is the only childbirth simulator which is actuated by electrical and pneumatic actuators. The aim of the second improvement is to synthesize control laws which improve the haptic rendering. The studied control laws allow the haptic interface to be controlled in position and its stiffness to be tuned in real time. This cartesian stiffness tuning is suitable to reproduce the behavior of the fetal head surrounded by the maternal pelvic muscles. The control law presented in this document have been obtained by the non-linear backstepping synthesis. Indeed, the non-linear phenomenons occurring in pneumatic cylinders can be taken into account by this method. The controller obtained in this work allows the BirthSIM end-effector to track a reference position but also allows the robot stiffness to be tuned in real time. The stiffness control is based on a gain tuning strategy. Finally, the performances have been evaluated in simulation and experimental tests. Those performances have also been compared to the other conventional compliant controllers of the robotic field.

Orientador: Marcelo Carvalho Minhoto Teixeira / Resumo: Esta tese propõe projetos de controle H_infinito e controle H_infinito chaveado para uma classe de sistemas lineares com incertezas invariantes no tempo, nos quais apenas a derivada do vetor de estado é considerada para realimentação. Neste cenário, as estratégias propostas utilizam dinâmicas auxiliares, cujas variáveis de estado estão disponíveis para realimentação, para controlar a planta original. São propostas estruturas e procedimentos de projeto por meio de desigualdades matriciais lineares. A primeira estrutura de controle apresentada, considera a saturação do atuador e um integrador para definir a estratégia de controle, e se as condições propostas são factíveis, assegura que o ponto de equilíbrio do sistema de malha fechada seja localmente assintoticamente estável, para todas as condições iniciais em uma região elipsoidal, contida em uma determinada região definida para a nova dinâmica. Na segunda estrutura, é proposta uma técnica sem o uso de integradores para definir as estratégias para o controle chaveado e controle H_infinito chaveado, para plantas com e sem saturação da entrada de controle. Na terceira estrutura, considera-se um caso mais geral porém com o uso de integradores para definir as estratégias para o controle chaveado e controle H_infinito chaveado, para plantas sem saturação da entrada de controle. Embora todos os projetos propostos considerem a dinâmica auxiliar, são asseguradas as propriedades de estabilidade, taxa de decaimento e custo garantido pa... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: This thesis proposes H_infinite switched control designs for a class of linear systems with time invariant uncertainties, where only the derivative of the state vector is considered for feedback. In this scenario, the proposed strategies use auxiliary dynamics, whose state variables is accessible for feedback, to control the original plant. The proposed structures and design procedures are based on LinearMatrix Inequalities (LMIs). The first structure considers the actuator saturation and an integrator to define the control strategy, and if the proposed conditions are feasible, it assures that the equilibrium point of the closed-loop system is locally asymptotically stable, for all initial conditions in an ellipsoidal region, which is within a given region defined for this new dynamic. In the second structure, we adopted a technique without the use of integrators to define the strategies for the switched control and H_infinite switched control, for plants with and without control input saturation. In the third structure, we study a more general case, however, with the use of integrators to define the strategies for the switched control and H_infinite switched control, for plants without control input saturation. Although all proposed designs consider the auxiliary dynamics, it is assured the stability, decay rate proprieties, and H_infinite guaranteed cost for the original plant. Simulation examples and practical implementation illustrate the effectiveness of the proposed app... (Complete abstract click electronic access below) / Doutor

Seit W. Diffie und M. Hellman im Jahr 1976 ihren Ansatz für einen sicheren kryptographischen Schlüsselaustausch vorgestellten, ist der sogenannte Diskrete Logarithmus zu einem zentrales Thema der Kryptoanalyse geworden. Dieser stellt eine Erweiterung des bekannten Logarithmus auf beliebige endliche Gruppen dar. In der vorliegenden Dissertation werden zwei von C. Diem eingeführte Algorithmen untersucht, mit deren Hilfe der diskrete Logarithmus in der Picardgruppe glatter, nichthyperelliptischer Kurven vom Geschlecht g > 3 bzw. g > 4 über endlichen Körpern berechnet werden kann. Beide Ansätze basieren auf der sogenannten Indexkalkül-Methode und benutzen zur Erzeugung der dafür benötigten Relationen spezielle Linearsysteme, welche durch Schneiden von ebenen Modellen der Kurve mit Geraden erzeugt werden. Um Aussagen zur Laufzeit der Algorithmen tätigen zu können, werden verschiedene Sätze über die Geometrie von Kurven bewiesen. Als zentrale Aussage wird zum einem gezeigt, dass ebene Modelle niedrigen Grades effizient berechnet werden können. Zum anderen wird bewiesen, dass sich bei genügend großem Grundkörper die Anzahl der vollständig über dem Grundkörper zerfallenden Geraden wie heuristisch erwartet verhällt. Für beide Aussagen werden dabei Familien von Kurven betrachtet und diese gelten daher uniform für alle glatten, nichthyperelliptischen Kurven eines festen Geschlechts. Die genannten Resultate führen schlussendlich zu dem Beweis einer erwarteten Laufzeit von O(q^(2-2/(g-1))) für den ersten der beiden Algorithmen, wobei q die Anzahl der Elemente im Grundkörper darstellt. Der zweite Algoritmus verbessert dies auf eine heuristische Laufzeit in O(q^(2-2/(g-2))), imdem er Divisoren von höherem Spezialiätsgrad erzeugt. Es wird bewiesen, dass dieser Ansatz für einen uniform gegen 1 konvergierenden Anteil an glatten, nichthyperelliptischen Kurven eines festen Geschlechts über Grundkörpern großer Charakteristik eine große Anzahl an Relationen erzeugt. Wiederum werden zum Beweis der zugrundeliegenden geometrischen Aussagen Familien von Kurven betrachtet, um so die Uniformität zu gewährleisten. Beide Algorithmen wurden zudem implementiert. Zum Abschluss der Arbeit werden die Ergebnisse der entsprechenden Experimente vorgestellt und eingeordnet.

We employ two efficient parallel approaches to reduce a model arising from a semi-discretization of a controlled heat transfer process for optimal cooling of a steel profile. Both algorithms are based on balanced truncation but differ in the numerical method that is used to solve two dual generalized Lyapunov equations, which is the major computational task. Experimental results on a cluster of Intel Xeon processors compare the efficacy of the parallel model reduction algorithms.

<p>The broad implementation of three-dimensional full-core modeling, with pin-resolved detail, for computational simulation and analysis of nuclear reactors highlights the importance of accuracy and efficiency in simulation codes for accurate and precise analysis. The primary objective of this dissertation is to develop a high-fidelity code capable of solving time-dependent neutron transport problems with 3D whole-core pin-resolved detail in nuclear reactor cores. Additionally, the dissertation explores the optimization of the code's parallelism to enhance its computational efficiency. To reduce the computational intensity associated with the direct 3D calculation of the neutron transport equation, a high-fidelity neutron transport code called PANDAS-MOC is developed using the 2D/1D approach. The 2D radial solution is obtained using the 2D Method of Characteristics (MOC), the axial 1D solution is determined through the Nodal Expansion Method (NEM), and then two solutions are coupled using transverse leakages to find the 3D solution. The convergence of the iterative scheme is accelerated using the multi-level coarse finite different mesh (ML-CMFD) technique. The code's validation and verification are carried out using the C5G7-TD benchmark exercises.</p> <p><br></p> <p>The significant and innovative aspect of this work involves parallelizing and optimizing the PANDAS-MOC code. Three parallel models are developed and evaluated based on the distributed memory and shared memory architecture: MPI parallel model (PMPI), Segment OpenMP threading hybrid model (SGP), and Whole-code OpenMP threading hybrid model (WCP). When computing the steady state of the C5G7 3D core with the same resources, the obtained speedup relationship between the three models is PMPI \(>\) WCP \(>\) SGP, whereas the WCP model only consumed 60\% of the memory of the PMPI model. Furthermore, the hybrid reduction in the ML-CMFD solver and the parallelism design of the MOC sweep are significant issues that decreased the speedup of WCP. Therefore, this study also addresses further optimizations of these two modules.</p> <p><br></p> <p>Concerning the MOC parallelism, two improvements are discussed: No-atomic schedule and Additional Axial Decomposition (AAD) parallelism. The No-atomic schedule evenly distributed the workload among threads and removes the \textit{omp atomic} clause from the code by predefining the MOC calculation sequence for each launched OpenMP thread while ensuring a thread-safe parallel environment. It can significantly reduce the calculation time and improve parallel efficiency. Furthermore, AAD divides the axial layers and OpenMP threads into multiple groups and restricts each thread to work on the layers designated to the same group. </p> <p>Meanwhile, Flag-Save-Update reduction is designed to increase the computational efficiency of the hybrid MPI/OpenMP reduction operations in the ML-CMFD module. It is accomplished by using the global arrays and status flags and establishing a tree configuration of all threads, and it includes no implicit and explicit barriers. In the case of the C5G7 3D core, the parallel efficiency of the MOC solver is about 0.872 when using 32 threads (=\#MPI \(\times\)\#OpenMP), and the Flag-Save-Update reduction yielded better speedup than the traditional hybrid MPI/OpenMP reduction, and its superiority is more obvious as more OpenMP threads are utilized. As a result, the WCP model outperforms the PMPI model for the overall steady-state calculation.</p> <p><br></p> <p>This research also investigates parallelizable preconditioners to accelerate the convergence of the generalized minimal residual method (GMRES) in the CMFD solver. Preconditioners such as Incomplete LU factorization (ILU), Symmetric Successive Over-relaxation (SOR), and Reduced Symmetric Successive Over-Relaxation (RSOR), are implemented in PANDAS-MOC. Except for RSOR, others are unsuitable for hybrid MPI/OpenMP parallel machines due to their inherent sequential nature and dependency on computation order. Their counterparts using the Red-Black ordering algorithm, namely RB-SOR, RB-RSOR, and RB-ILU, are formatted and examined on benchmark reactors such as TWIGL-2D, C5G7-2D, C5G7-3D, and their corresponding subplane models (TWIGL-2D(5S), C5G7-2D(5S), C5G7-3D(5S)), with relaxed convergence criteria (\(10^{-3}\)). Results show that all preconditioners significantly reduce the required number of iterations to converge the GMRES solutions, and RB-SOR is the best one for most reactors. In the case of C5G7-3D(5S), preconditioners exhibit similar sublinear speedup but demonstrate varying runtimes across all tests for both MG-GMRES and 1G-GMRES. However, the speedup results in 1G-GMRES are more than twice as high as those in MG-GMRES. RB-RSOR has an optimal efficiency of 0.6967 at (4,8), while RB-SOR and RB-ILU have optimal efficiencies of 0.6855 and 0.7275 at (32,1), respectively.</p>

The suspension system in vehicles supports the vehicle's road stability and ride quality by scaling down the vibration responses resulting from road surface's roughness. This research focuses on fatigue life analysis of coil spring component. Static linear analysis is conducted on the 3D model of helical coil spring to investigate deformation and stress responses. Modal analysis evaluates the characteristics of vibration, i.e. natural resonance frequencies and corresponding mode shapes. The stress frequency response is generated after performing the harmonic analysis on the spring. Dynamics and performance of spring are analyzed over practical frequency range of 0 Hz to 200 Hz. Fatigue life estimation of vehicle suspension spring is performed using the stress data obtained from frequency response analysis. The stress-life (S-N) approach is utilized for fatigue life assessment of suspension spring. This durability analysis technique can be utilized in the automotive industry to improve reliability of vehicles. The outcome of this research can contribute in analysis and design of modern smart vehicles. / Master of Science / The suspension system in vehicles supports the vehicle's road stability and ride quality by scaling down the vibration responses resulting from road surface's roughness. This research focuses on the fatigue life analysis of suspension spring component. Initial phase of analysis is conducted to investigate the deformation and stress in 3D model of spring. Dynamics and performance of spring are analyzed over applicable frequency range of 0 Hz to 200 Hz. Fatigue life of vehicle suspension spring is evaluated using stress data from frequency response analysis. This durability analysis technique can be utilized in the automotive industry to improve reliability of vehicles. The outcome of this research can contribute in analysis and design of modern smart vehicles.

[pt] Esta dissertação apresenta uma comparação entre técnicas de discretização de controladores, considerando diferentes estruturas de controle. Os tipos de sistemas estudados neste trabalho de pesquisa serão sistemas lineares, invariantes no tempo, determinísticos, causais e monovariáveis. O desempenho das técnicas de discretização serão comparados via figuras de mérito tradicionais, considerando os métodos de discretização, as estruturas dos controladores e os tipos de planta habituais (incluindo problemas benchmarch), sob especificações dadas quanto aos regimes permanente e transiente. / [en] This essay proposes a comparison between techniques of controllers´ discretization considering different controlling structures. The types of systems studied in this research will be linear systems, time-invariant, deterministic, casual and single-variable. The performance of discretization techniques will be compared through figures of traditional aptitude, considering the discretization methods, the controller structures and the kinds of plants (including - benchmarch - problems), under given specifications according to permanent and transitory systems. / [es] Esta disertación presenta una comparación entre técnicas de discretización de controladores, considerando diferentes extructuras de control. Los tipos de sistemas estudiados en este trabajo de investigación son sistemas lineales, invariantes en el tiempo, determinísticos, causales y univariados. Se compara el desempeño de las técnicas de discretización utilizando figuras de mérito tradicionales, considerando los métodos de discretización, las extructuras de los controladores y los tipos de planta habituales (incluyendo problemas - benchmarch - ), bajo especificaciones dadasen relación a los régimenes permanente y transiente.