Diversos problemas envolvem a classificação de dados em categorias, também denominadas classes. A partir de um conjunto de dados cujas classes são conhecidas, algoritmos de Aprendizado de Máquina (AM) podem ser utilizados na indução de um classificador capaz de predizer a classe de novos dados do mesmo domínio, realizando assim a discriminação desejada. Dentre as diversas técnicas de AM utilizadas em problemas de classificação, as Máquinas de Vetores de Suporte (Support Vector Machines - SVMs) se destacam por sua boa capacidade de generalização. Elas são originalmente concebidas para a solução de problemas com apenas duas classes, também denominados binários. Entretanto, diversos problemas requerem a discriminação dos dados em mais que duas categorias ou classes. Nesta Tese são investigadas e propostas estratégias para a generalização das SVMs para problemas com mais que duas classes, intitulados multiclasses. O foco deste trabalho é em estratégias que decompõem o problema multiclasses original em múltiplos subproblemas binários, cujas saídas são então combinadas na obtenção da classificação final. As estratégias propostas visam investigar a adaptação das decomposições a cada aplicação considerada, a partir de informações do desempenho obtido em sua solução ou extraídas de seus dados. Os algoritmos implementados foram avaliados em conjuntos de dados gerais e em aplicações reais da área de Bioinformática. Os resultados obtidos abrem várias possibilidades de pesquisas futuras. Entre os benefícios verificados tem-se a obtenção de decomposições mais simples, que requerem menos classificadores binários na solução multiclasses. / Several problems involve the classification of data into categories, also called classes. Given a dataset containing data whose classes are known, Machine Learning (ML) algorithms can be employed for the induction of a classifier able to predict the class of new data from the same domain, thus performing the desired discrimination. Among the several ML techniques applied to classification problems, the Support Vector Machines (SVMs) are known by their high generalization ability. They are originally conceived for the solution of problems with only two classes, also named binary problems. However, several problems require the discrimination of examples into more than two categories or classes. This thesis investigates and proposes strategies for the generalization of SVMs to problems with more than two classes, known as multiclass problems. The focus of this work is on strategies that decompose the original multiclass problem into multiple binary subtasks, whose outputs are then combined to obtain the final classification. The proposed strategies aim to investigate the adaptation of the decompositions for each multiclass application considered, using information of the performance obtained for its solution or extracted from its examples. The implemented algorithms were evaluated on general datasets and on real applications from the Bioinformatics domain. The results obtained open possibilities of many future work. Among the benefits observed is the obtainment of simpler decompositions, which require less binary classifiers in the multiclass solution.

Orientador: Robson Pederiva / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecânica / Made available in DSpace on 2018-08-19T20:21:22Z (GMT). No. of bitstreams: 1 Silva_ViniciusAugustoDiniz_M.pdf: 16029999 bytes, checksum: a3585cbd021c6f84637d409a34a51962 (MD5) Previous issue date: 2012 / Resumo: Motores elétricos são componentes essenciais na grande maioria dos processos industriais. As diversas falhas nas máquinas de indução podem gerar consequências drásticas para um processo industrial. Os principais problemas estão relacionados ao aumento dos custos, piora nas condições do processo e de segurança e qualidade do produto final. Muitas destas falhas mostram-se progressivas. Neste trabalho, apresenta-se uma contribuição ao estudo de métodos de detecção de falhas em motores elétricos usando Máquinas de Vetores de Suporte (SVM), treinadas a partir de sinais de vibração obtidos experimentalmente. A metodologia desenvolvida é usada para classificar a excitação devido a falhas mecânicas e elétricas, além da condição normal de funcionamento, utilizando apenas um sensor de vibração. Através da seleção de parâmetros é possível reduzir o número de entradas capazes de representar os sinais utilizados para o treinamento das SVMs. A normalização proposta permitiu melhorar as taxas de acerto, quando se quer classificar falhas em diferentes níveis de severidade das que foram utilizadas para o treinamento. Os resultados mostraram que a metodologia apresentada pode ser adaptada para ser utilizada em aplicações práticas industriais e poderá ser no futuro uma saída viável para uma manutenção industrial eficiente e eficaz / Abstract: Electric motors are essential components in most industrial processes. The several faults in induction machines can produce drastic consequences for an industrial process. The main problems are related to rising costs, decrease conditions in the process and safety and quality of the final product. Many of these failures are progressive. In this paper, we present a contribution to the study of methods for detecting faults in induction motors using Support Vector Machines (SVM) trained from vibration signals obtained experimentally. The developed methodology is used to classify the excitation due to mechanical and electrical failures, in addition to normal operating condition, using only a vibration sensor. Through the feature selection is possible to reduce the number of inputs that represent the signals used for training the SVMs. The proposed standardization has improved the accuracy rates when we want to classify failures at different levels of severity that were used for training. The results showed that this methodology can be adapted for use in industrial and practical applications and in the future may be a viable approach to an efficient and effective industrial maintenance / Mestrado / Mecanica dos Sólidos e Projeto Mecanico / Mestre em Engenharia Mecânica

L’estimation de la pose du corps humain est un problème difficile en vision par ordinateur et les actions de toutes les difficultés de détection d’objet. Cette thèse se concentre sur les problèmes de l’estimation de la pose du corps humain dans les images ou vidéo, y compris la diversité des apparences, les changements de scène et l’éclairage de fond de confusion encombrement. Pour résoudre ces problèmes, nous construisons un modèle robuste comprenant les éléments suivants. Tout d’abord, les méthodes top-down et bottom-up sont combinés à l’estimation pose humaine. Nous étendons le modèle structure picturale (PS) de coopérer avec filtre à particules recuit (APF) pour robuste multi-vues estimation de la pose. Deuxièmement, nous proposons plusieurs parties de mélange à base (MMP) modèle d’une partie supérieure du corps pour l’estimation de la pose qui contient deux étapes. Dans la phase de pré-estimation, il y a trois étapes: la détection du haut du corps, catégorie estimation du modèle pour le haut du corps, et la sélection de modèle complet pour pose estimation. Dans l’étape de l’estimation, nous abordons le problème d’une variété de poses et les activités humaines. Enfin, le réseau de neurones à convolution (CNN) est introduit pour l’estimation de la pose. Un Local Multi-résolution réseau de neurones à convolution (LMR-CNN) est proposé pour apprendre la représentation pour chaque partie du corps. En outre, un modèle hiérarchique sur la base LMR-CNN est défini pour faire face à la complexité structurelle des parties de branche. Les résultats expérimentaux démontrent l’efficacité du modèle proposé / Human pose estimation is a challenging problem in computer vision and shares all the difficulties of object detection. This thesis focuses on the problems of human pose estimation in still images or video, including the diversity of appearances, changes in scene illumination and confounding background clutter. To tackle these problems, we build a robust model consisting of the following components. First, the top-down and bottom-up methods are combined to estimation human pose. We extend the Pictorial Structure (PS) model to cooperate with annealed particle filter (APF) for robust multi-view pose estimation. Second, we propose an upper body based multiple mixture parts (MMP) model for human pose estimation that contains two stages. In the pre-estimation stage, there are three steps: upper body detection, model category estimation for upper body, and full model selection for pose estimation. In the estimation stage, we address the problem of a variety of human poses and activities. Finally, a Deep Convolutional Neural Network (DCNN) is introduced for human pose estimation. A Local Multi-Resolution Convolutional Neural Network (LMR-CNN) is proposed to learn the representation for each body part. Moreover, a LMR-CNN based hierarchical model is defined to meet the structural complexity of limb parts. The experimental results demonstrate the effectiveness of the proposed model

L’objectif principal de cette thèse consiste à mettre en évidence la persistance du capitalisme familial en Suisse au cours du 20e siècle, et sa résistance aux capitalismes managérial et financier qui sont censés lui avoir succédé. Pour ce faire, nous avons retenu vingt-deux grandes entreprises du secteur des machines, de l’électrotechnique et de la métallurgie – principale branche de l’industrie suisse pour la période considérée –, pour lesquelles ont été recensés les membres des conseils d’administration et les principaux dirigeants exécutifs pour cinq dates-repère couvrant le siècle (1910, 1937, 1957, 1980 et 2000). Cette thèse s’inscrit dans une démarche pluridisciplinaire qui relève à la fois de l’histoire d’entreprise et de la sociologie des dirigeants, et fait appel à différentes méthodes telles que l’analyse de réseau et l’analyse prosopographique. Elle s’articule autour de trois axes de recherche principaux : le premier vise à mettre en évidence l’évolution des modes de gouvernance dans notre groupe d’entreprises, le second investit la question de la coordination patronale et le troisième a pour but de dresser un portrait collectif des élites à la tête de nos vingt-deux firmes. Nos résultats montrent que durant la majeure partie du siècle, la plupart de nos entreprises sont contrôlées par des familles et fonctionnent sur un mode de coordination hors marché qui repose notamment sur un réseau dense de liens interfirmes, le profil des dirigeants restant dans l’ensemble stable. Si la fin du siècle est marquée par plusieurs changements qui confirment l’avènement d’un capitalisme dit financier ou actionnarial et la mise en place de pratiques plus concurrentielles parmi les firmes et les élites industrielles, le maintien du contrôle familial dans plusieurs entreprises et la persistance de certains anciens mécanismes de coopération nous incitent cependant à nuancer ce constat. / The main objective of this research is to highlight the persistence of family capitalism in Switzerland during the 20th century and its resistance to managerial and financial capitalisms that succeeded. For this purpose, we focus on twenty-two big companies of the machine, electrotechnical and metallurgy sector – the main branch of the Swiss industry for the considered period – whose boards of directors and executive managers have been identified for five benchmarks across the century (1910, 1937, 1957, 1980 and 2000). This thesis relates to business history and elites sociology, and uses different methods such as network analysis and prosopography. It is articulated around three main parts. The aim of the first one is to identify the evolution of corporate governance in our twenty-two enterprises, the second part concentrates on interfirms coordination and the objective of the last one is to highlight the profile of the corporate elite leading our firms. Our results show that during the main part of the century, most of the companies were controlled by families and were characterized by non-market mechanisms of coordination such as interlocking directorates ; moreover, the profile of the corporate elite remained very stable. Although some major changes that took place by the end of the century confirmed a transition towards financial capitalism and more competitive interaction among firms and the corporate elite, the persistence of family control in several companies and the maintaining of some former mechanisms of coordination allow us to put this evolution into perspective.

Der ständig steigende Bedarf an Lebensmitteln und Getränken insbesondere in Schwellenländern wie China, Indien und Brasilien fordert von der Verpackungsindustrie eine Leistungssteigerung in gleichem Maße. Um den Durchsatz einer einzelnen Verpackungs- und Füllmaschine unter Beibehaltung von Verpackungsqualität und -form zu steigern, gibt es grundsätzlich zwei Möglichkeiten. Die eine Verbreiterung der Maschine nach sich ziehende Erhöhung der Anzahl der gleichzeitig bearbeiteten Verpackungen, sowie die weitestgehend bauraumneutrale Verringerung der Taktzeit. Letzteres wurde von der SIG Combibloc Systems GmbH, einem der weltweit führenden Systemanbieter von Packstoff und entsprechenden Füllmaschinen für Getränke und Lebensmittel, für eines ihrer Verpackungssysteme in Angriff genommen. Ziel war die Erhöhung des Durchsatzes um 50 % von 24.000 auf 36.000 Verpackungen pro Stunde. [... aus der Einleitung]

The subject of this PhD thesis is the development of an efficient system that can simultaneously pump and separate a gas-liquid mixture, in particular an oil-air mixture. Two-phase flows are encountered in many applications (petroleum extraction, flow in nuclear power plant pumps, pulp and paper processing, etc.) but this study is mainly focused on lubrication systems of aircraft gas turbine engines.<p><p>The pump and separator system (PASS) for two-phase flows developed in this PhD thesis aims to perform three functions simultaneously:<p>• Send back the oil to the tank (oil pumping)<p>• Separate the air from the oil (de-aeration)<p>• Separate the oil from the air (de-oiling) and release the sealing air into the atmosphere (venting). <p>Particular care is given to the liquid flow rate lost at the gas outlet of the system.<p>Consequently, it could replace the scavenge pumps and oil-air separators existing in present lubrication systems. This modification provides several advantages: simplification of the lubrication circuit, reduction of oil consumption and of the size of the lubrication system.<p><p>This research is divided into three axes: the theoretical study of the important physical mechanisms taking place inside the two-phase flow pump and separator system, the experimental development, tests and optimization of different PASS prototypes, and also the numerical simulations of the two-phase flow inside these prototypes. Although the experiments were the central pillar of this research, the three axes were closely imbricated.<p><p>The PASS design includes three main components:<p>• An inlet chamber with one or several tangential inlets giving a natural centrifugation to the flow,<p>• An impeller (forced centrifugation) with an axial and a radial part followed by a volute chamber,<p>• A metallic foam that lets pass micron and sub-micron droplets and which is followed by an axial vent port.<p><p>The centrifugation causes the liquid (oil) to move radially outwards in an annular body (a liquid ring) generating pressure. The thickness of this liquid ring inside the impeller is mainly determined by the pressure coefficient (related to the back-pressure and the rotational speed). When the back-pressure increases, the thickness of the liquid ring increases too. An advantage of the PASS is that it does not impose any relation between the liquid head and the liquid flow rate, contrary to common centrifugal pump. It self-regulates the radial position of the gas-liquid interface to sustain the operating conditions.<p><p>The de-aeration efficiency mainly depends on the pressure coefficient (for a constant liquid viscosity or temperature) or on the thickness of the liquid ring. The pressure gradient which appears in the liquid rotating in an annular body acts like a dam for the gas phase. Indeed, the gas movement is mainly determined by the pressure field (buoyancy) while the liquid distribution is dominated by centrifugal and Coriolis forces. Buoyancy tends to accumulate the gas phase near low pressure areas (PASS hub, suction side of the blades, clearances between closed impeller and casing).<p><p>The first oil-air PASS prototype produces high viscous losses due to the high peripheral velocity and liquid viscosity. Therefore, the pumping efficiency is poor compared to common impeller pumps. However, the pumping is not the key function of the PASS and a power consumption below 5 kW is acceptable for the application considered in this work. For applications that require lower power consumptions, a reduction of the rotational speed must be considered.<p><p>Thus, the rotational speed and the impeller diameter are two major constraints for the PASS design which determine the de-aeration and pumping efficiencies. The impeller diameter also influences the size of passage sections for the air flow. The air velocity must be kept as low as possible because the entrainment of droplets increases when the air velocity rises (drag forces on droplets). Indeed, this large influence of the air flow rate on the oil consumption (de-oiling efficiency) was demonstrated by a theoretical analysis, the experiments and the CFD simulations. The production of droplets in the inlet pipes when the two-phase flow is annular is a key phenomenon regarding the oil consumption.<p><p>In addition to the air flow rate, other variables also influence the oil consumption:<p>• Air-oil temperature: when the temperature rises, the oil consumption increases because the surface tension and the oil density are reduced. Moreover, as the air density also decreases, the air velocity rises.<p>• Oil flow rate: the oil consumption rises more or less linearly with the oil flow rate. However, the influence of the oil flow rate on the inlet droplet size is uncertain.<p>• Rotational speed: the rotational speed has obviously a strong impact on the oil consumption without metallic foam. However, experiments showed that the metallic foam efficiency is almost independent on the rotational speed. Therefore, the oil consumption with the Retimet foam does not depend on the PASS rotational speed.<p>• Altitude or air density: the oil consumption decreases when the air density is reduced because the drag forces on droplets also decrease.<p>The gas density (altitude) is also supposed to influence the de-aeration efficiency but this could not be tested or simulated in this work (the de-aeration efficiency gets probably better when decreasing the gas density because the buoyancy forces increase).<p><p>Theory, experiments and numerical simulations also allowed the prediction of performance of the first oil-air prototype for real in-flight operating conditions. Two problems have been identified: the de-aeration efficiency at MTO and cruise ratings and the oil leak throughout the vent in cold start and windmilling. To solve them, some modifications of the lubrication system have been suggested. With these modifications, the oil-air PASS should become very efficient and attractive for engine manufacturers. / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished

The aim of the thesis is to propose active damping as a potential control strategy for chatter instability in machine tools.<p>The regenerative process theory explains chatter as a closed loop interaction between the structural dynamics and the cutting process. This is considered to be the most dominant reason behind machine tool chatter although other instability causing mechanisms exist.<p>The stability lobe diagram provides a quantitative idea of the limits of stable machining in terms of two physical parameters: the width of contact between tool and the workpiece, called the width of cut and the speed of rotation of the spindle. It is found that the minimum value of the stability limit is proportional to the structural damping ratio for turning operations. This important finding provides the motivation of influencing the structural dynamics by active damping to enhance stability limits of a machining operation.<p>A direct implementation of active damping in an industrial environment may be difficult. So an intermediate step of testing the strategy in a laboratory setup, without conducting real cutting is proposed. Two mechatronic "Hardware in the Loop" simulators for chatter in turning and milling are presented, which simulate regenerative chatter experimentally without conducting real cutting tests. A simple cantilever beam, representing the MDOF dynamics of<p>the machine tool structure constitutes the basic hardware part and the cutting process is simulated in real time on a DSP board. The values of the cutting parameters such as spindle speed and the axial width of cut can be changed on the DSP board and the closed loop interaction between the structure and the cutting process can be led to instability.<p><p>The demonstrators are then used as test beds to investigate the efficiency of active damping, as a potential chatter stabilization strategy. Active damping is easy to implement, robust and does not require a very detailed model of the structure for proper functioning, provided a collocated sensor and actuator configuration is followed. The idea of active damping is currently being implemented in the industry in various metal cutting machines as part of the European Union funded SMARTOOL project (www.smartool.org), intended to propose smart chatter control technologies in machining operations. / Doctorat en sciences appliquées / info:eu-repo/semantics/nonPublished

Ces vingt dernières années, le domaine du design et de la synthèse de machines moléculaires complexes a fait d’énormes progrès,souvent inspiré par la beauté de la machinerie présente dans les systèmes vivants. Cependant, l’amplification des mouvements d’un grand nombre de machines moléculaires à des échelles de tailles largement supérieures à leurs dimensions restent un défi théorique et expérimental ambitieux et ardu. Ce travail décrit comment l’organisation de machines et moteurs moléculaires dans des réseaux polymères supramoléculaires ou covalents permet de synthétiser des matériaux dans lesquels leurs mouvements individuels nanométriques sont amplifiés jusqu’à l’échelle macroscopique. Les trois premiers chapitres décrivent l’utilisation d’une architecture de type [c2]daisy chains, une molécule capable d’effectuer des contraction/extensions similaires aux mouvements des sarcomères présents dans les muscles, dans des réseaux polymères supramoléculaires et covalents. Leur introduction dans des polymères supramoléculaires à liaisons hydrogène basées sur le motif de reconnaissance uracil:2,6-diacetylaminopyridine associé à des interactions latérales tel que les interactions π résulta en la formation de fibre supramoléculaires contractiles dont la taille et la morphologie a pu être commuté entre deux états étendus et contractés. L’incorporation de motifs uréidopyrimidinone comme connecteur supramoléculaire en revanche donna accès à des gels supramoléculaires, évoluant vers un état liquide lors de la contraction des chaines polymères. Finalement, l’inclusion de daisy chains dans un réseau polymère 3D a donné accès à un gel chimique. Ce matériau a pu être contracté et étendu à l’échelle macroscopique grâce à l’action combinée des machines moléculaires le constituant.Le quatrième chapitre est dédié à l’amélioration d’un gel contractile basé sur l’utilisation de moteurs moléculaires rotatif comme noeud de réticulation d’un réseau polymère. Une unité modulatrice, capable d’être commuté entre un état “ouvert” et “fermé”, a été introduite dans le réseau aux côtés du moteur. Le modulateur dans son état “fermé” permet aux moteurs moléculaires de contracter efficacement le réseau, tandis que dans son état ouvert il permet aux chaines de se dérouler alors que le moteur ne tourne pas, ce qui provoque l’extension du réseau qui retourne à sa taille initiale. En résumé, le travail décrit dans ce manuscrit illustre que des machines moléculaires soigneusement conçue peuvent être introduites dans des réseaux polymères, fournissant des matériaux dont les propriétés macroscopiques sont affectées par les mouvements nanoscopiques de ses constituants. Ces résultats fournissent des pistes et une base fondamentale pour l’élaboration d’une nouvelle classe de matériaux contractiles basés sur des machines moléculaires. / The last twenty years have seen tremendous progresses in the design and synthesis of complex molecular machines, often inspired by the beauty of the machinery found in biological systems. However, amplification of the molecular machines motion over several orders of magnitude above their typical length scale is still an ambitious challenge. This work describes how self-organization of molecular machines or motors allows for the synthesis of materials translating the motions of their components into a macroscopic response. The three first chapters describe the use of a [c2]daisy chains architecture, a molecule able to perform contraction/extension motions similarly to the sarcomere units of muscles, into systems such as supramolecular polymers and covalent networks. Their inclusion into hydrogen bonding supramolecular polymers based on the uracil:2,6-diacetylaminopyridine recognition motifs combined with lateral interactions such as π-stacking provided micrometric muscle-like fibers contracting and extending upon deprotonation and protonation.The incorporation of ureidopyrimidone moieties as supramolecular connectors yielded highly organized gels, which evolved to a liquidstate upon contraction of the polymer chains. Finally, covalent poly[c2]daisy chains were synthesized and investigated, notably the formation of a 3D network swelling into a gel. This material could contract and extend at the macroscopic scale upon contraction and extension of the molecular machines used as monomers. Finally, a fourth chapter is dedicated to the improvement of contractile chemical gels made by using a molecular motor as reticulating nodes. A modulating unit, able to be switched between a “closed” and an “opened” state, was introduced into the polymer network along with the motor. The locked structure in the “closed” state allowed contraction of the gel upon rotation of the molecular motors, while the “opened” state allowed unwinding of the entangled polymer chains and extension of the gel when the motor is off. Overall, the work presented in this manuscript demonstrates that carefully designed molecular machines can be incorporated into large supramolecular or covalent assemblies, providing materials which collective motions alter their macroscopic properties. These results provide valuable insights for the elaboration of a new class of muscle-like materials based on molecular machines.

Le moteur à réluctance variable (MRV) est en train de gagner beaucoup d'intérêt dans les applications de traction électrique grâce à la compatibilité significative entre ses caractéristiques naturelles et les exigences de l'industrie automobile. Cependant, à cause de la caractéristique fortement non linéaire des grandeurs électromagnétiques du MRV, l'usure du moteur, les variations paramétriques dues aux tolérances de fabrication, et la dérive des paramètres durant le fonctionnement, les stratégies de commande classique ne permettent pas d'assurer des hautes performances de suivi de trajectoire et de robustesse en stabilité dans l'environnement de fonctionnement d'un véhicule électrique. Dans ce contexte, le premier objectif de cette thèse est de proposer des stratégies de commande robustes du MRV en prenant en considération les contraintes imposées par les applications de traction. Afin d'atteindre cet objectif, une structure de commande en cascade, composée d'une boucle externe de vitesse et d'une boucle interne de courant, est envisagée. Les objectifs de synthèse sont ensuite fixés à travers un cahier des charges formulé en termes de gabarits fréquentiels, et les correcteurs robustes et adaptatifs de vitesse et de courant sont ainsi synthétisés en se basant sur les approches de commande H-infini et LPV/H- infini. La faisabilité des correcteurs constitue également un critère important dans les applications de traction électrique où la puissance de calcul embarqué dans le véhicule est très limitée. Ce critère est pris en considération dans cette thèse à travers l'utilisation des approches de synthèse des correcteurs H- infini d'ordre fixe. Les performances de ces correcteurs sont analysées et comparées avec les performances des correcteurs H- infini classiques à travers une étude par la μ-analyse de la robustesse en stabilité et une évaluation expérimentale. Le deuxième objectif de ces travaux de thèse est d'évaluer l'intérêt de la commande sans capteurs mécanique du MRV en se basant sur des méthodes d'estimation. Des observateurs robustes basés sur les modes glissants et le filtrage de Kalman sont ainsi synthétisés. Les performances et la robustesse de ces observateurs sont ensuite comparées expérimentalement dans une perspective d'une application de traction électrique. / The switched reluctance motor (SRM) is gaining much interest in electric vehicle applications. It includes robustness to harsh operational conditions, rugged structure, fault tolerant operation and a wide range of speed. However, due to the highly nonlinear electromagnetic characteristic, wear and tear of motor, manufacturing tolerances and parameters drift during operation, classical control strategies do not ensure high dynamical performance and robustness under various operating conditions in electric vehicle environment. In this context, the first purpose of this thesis is to propose robust control strategies of the SRM taking into account the constraints imposed by electric vehicle applications. To achieve this objective, a cascade control scheme is adopted; it consists of an outer speed loop and an inner current loop. Controllers design aims are fixed through specifications formulated in frequency domain, robust and adaptive controllers of speed and currents are thus synthesized based on H- infini and LPV/H- infini control approaches. Feasibility of these controllers is also an important criterion in electric vehicles applications where embedded computing power is very limited. This criterion is taken into account in this thesis by using fixed order controller synthesis approaches. The performances of these controllers are analyzed and compared with the performances of standard H- infini controllers through an experimental evaluation and a robustness analysis is performed using the concept of the Singular Structured Value, i.e. the μ-analysis. The second objective of this thesis is to assess the interest of the sensorless control of SRM. Robust observers based on second order sliding mode (SMO2) and Kalman filtering techniques (EKF) are designed to estimate the mechanical states of SRM. Performances and robustness of these observers are then compared experimentally in perspective of electric vehicle application.

Le travail de thèse s’inscrit dans le cadre du projet TRAX. Il s’agit là du développement de moteurs électriques destinés à la traction des petits véhicules électriques urbains. Les caractéristiques clés d’une machine électrique pour une application de traction sont le couple, le rendement, la fiabilité, l’encombrement et la plage de vitesse à puissance maximale (dé-fluxage). Les machines électriques à aimants permanents répondent à ces exigences. C’est pourquoi ce travail de thèse s’est intéressé à l’étude des performances de machines à aimants permanents en proposant une étude comparative. Un état de l’art basé sur l’étude des brevets des machines électriques dans les applications automobiles est réalisé. Une description des dernières avancées des moteurs électriques principalement des structures à aimants permanents est fournit. Nous avons modélisé une première machine avec une méthode ana-lytique simplifiée basée sur la résolution des équations de Maxwell en 2D. Cette méthode est confrontée à une méthode numérique. Trois nouvelles machines sont définies à partir de la première en modifiant la configuration du rotor. La comparaison de quatre structures à aimants permanents est réalisée par des modèles numériques. Les performances à vide et en charge ainsi que le calcul des inductances dans l’axe direct et en quadrature sont évaluées. Les résultats théoriques sont comparés aux essais expérimentaux. / The thesis is part of the TRAX project. It deals with development of electric motors used for traction of small urban electric vehicles. The key characteristics of an electric machine for traction application are the torque, efficiency, reliability, size and flux-weakening. The permanents magnets electric machines meet these requirements. That’s why this thesis takes interest in the performances of permanents magnets machines by proposing a comparative study. A study of patents for electrical machines in automotive applications is realized. A description of the latest advances in electrical motors, mainly in permanent magnet structures, is provided. We started to model a first machine with a simplified analytical method based on the resolution of Maxwell's equations in 2D. This method is compared with a numerical method. Three new machines are defined from the first one by changing the configuration of the rotor. The comparison of four structures with permanent magnets is realized by numerical models. No load and load performances, as well as the calculation of inductances in the direct and quadrature axis, are evaluated. The theoretical results are compared with experimental tests.