No description available.

Ce travail à la fois expérimental et numérique s'inscrit dans le cadre des études menées sur le captage des polluants sur les machines tournantes et porte plus particulièrement sur la caractérisation de la dispersion des fines particules d'usinage : i.e. ces particules sont assimilées à un traceur passif du fait de leur temps de réponse aérodynamique négligeable. L'objectif recherché à travers cette étude est de développer une méthode numérique de conception des captages des polluants sur machines d'usinage. Le premier travail a donc consisté à réaliser un dispositif expérimental recréant une situation similaire à une opération de meulage, mais dans un cadre parfaitement contrôlé. L'écoulement retenu est celui engendré par un cylindre en rotation dans une veine d'essai ventilée. Un polluant particulaire, représenté par des microbilles de verre sphériques, est généré dans la veine d'essai (en utilisant un système d'ensemencement en particules), recréant ainsi les effets aérauliques des plus grosses particules d'usinage. Un gaz traceur, représentant la fraction fine de particules, est injecté simultanément avec le jet de particules via un capillaire placé à la source. Les propriétés du jet de particules obtenu sont caractérisées par vélocimétrie par suivi de particules (PTV) afin d'obtenir des données d'entrée et de validation pour les simulations numériques. La méthode de suivi de particules, mise au point dans le cadre de cette étude reste robuste même dans les zones fortement chargées en particules (i.e. la région source du jet). Le champ de vitesse de la phase gazeuse a quant à lui été caractérisé par vélocimétrie par images de particules (PIV), au moyen d'un code développé spécifiquement (Belut 2006 [4]). La dispersion d'un gaz traceur (hexafluorure de soufre - SF6) émis simultanément avec le jet de particules est ensuite étudiée expérimentalement : la dispersion d'un tel gaz est en effet jugée représentative de celle des plus fines particules d'usinage dont le temps de réponse aérodynamique est négligeable. Enfin, une modélisation complète du banc d'essai est réalisée afin de permettre une comparaison avec les résultats expérimentaux pour ainsi progresser dans la validation des modèles utilisés pour décrire la dispersion d'un traceur gazeux représentatif des fines particules. Le logiciel commercial Fluent est employé pour les simulations numériques / This study based at the same time in experimental and numerical aspect, is part of the framework to understand deeply the capture of pollutants on rotating machines and focuses especially on the characterization of the dispersion of fine particles generated by machining operations : i.e. these particles are assimilated as a passive tracer due to their negligible aerodynamic response time. The main objective covered by this study is to develop a numerical method of close capture exhausts systems for machining devices. The first task was to achieve an experimental device re-creating a similar grinding operation in a perfectly controlled environment. The flow used is defined as the one generated by a rotating cylinder in a ventilated test rig. A pollutant of particles, represented by spherical glass beads, is produced inside the test rig (by using a seeding system of particles), recreating therefore the aerodynamic effects of largest machining particles. A tracer gas, representing the fraction of fine particles, is injected simultaneously with the jet of particles through a capillary tube placed at the source. The properties of the jet of particles obtained are characterized by particle tracking velocimetry (PTV) allowing then to obtain data input and validation of numerical simulations. This particle tracking technic, developed in this study remains as robust method even in heavily loaded particles (i.e. the source region of the jet). The velocity field of the gas phase has meanwhile been characterized by particle image velocimetry (PIV), using a code developed specifically for this assay (Belut 2006 [4]). The dispersion of a tracer gas (sulfur hexafluoride - SF6) emitted simultaneously with the jet of particles is then studied experimentally : the dispersion of such a gas is indeed considered representative of finest particles which aerodynamic response time is negligible. Finally, a complete modeling of the experimental test rig is performed to allow comparison with experimental results in order to progress in the validation of models used to describe the dispersion of a tracer gas representative of finest particles. The commercial soft ware FLUENT is used for numerical simulations

Cette thèse porte sur l’étude et l’optimisation aérothermique d’un alterno-démarreur utilisé dans les véhicules hybrides. Ces machines produisant beaucoup plus de puissance qu’un alternateur classique, leur refroidissement est donc critique. La machine est modélisée en utilisant la méthode nodale en régime permanent qui utilise des réseaux de conductances thermiques. Pour alimenter le modèle, on utilise des corrélations issues de la littérature pour modéliser les transferts convectifs et on effectue des calculs CFD de la machine complète pour obtenir la répartition des débits. Les résultats obtenus numériquement sont ensuite validés expérimentalement à l’aide d’essais par Vélocimétrie par Images de Particules et d’essais thermiques par mesure thermocouples. Dans un deuxième temps, on couple un algorithme d’optimisation au code pour obtenir une géométrie de la machine optimale d’un point de vue thermique. La méthode retenue est l’Optimisation par Essaim Particulaire (PSO). L’optimisation se fait sur la taille des têtes de bobines, la position des ventilateurs et la section des canaux rotoriques. On obtient des géométries différentes selon les objectifs que l’on cherche à atteindre. La dernière partie de la thèse porte sur l’optimisation multi-objectifs d’un dissipateur située sur la partie électronique à l’arrière de l’alternateur : le dissipateur doit refroidir le plus possible l’électronique sans pour autant perturber l’écoulement. On étudie aussi plusieurs formes d’ailettes pour atteindre ces objectifs. / The goal of this thesis is the aero-thermal study and optimization of a starter-alternator used in hybrid cars. This kind of machines being more powerful than a regular alternator, their cooling is critical. The machine is modeled using lumped method in steady state which uses networks of thermal conductances. The inputs for the model are obtained using correlations from bibliography for the convective heat transfers and three dimensional CFD for the flow rates inside the machine. The numerical results are validated by experimental results with PIV for the fluid results and a machine fitted with thermocouples for the thermal part. In the second part, the thermal code is coupled with an optimization algorithm to obtain an optimal geometry of the machine from a thermal point of view. The method chosen is Particle Swarm Optimization (PSO). The parameters are the sizes of the end-windings, the positions of the fans and the cross section of the rotor channels. For different objectives, different optimal geometries are obtained. The last part of this work aims at the multi-objectives optimization of a heat sink located at the back of the machine. The heat sink has to be thermally efficient but should not affect the flow. Different shapes of fins are also studied.

Bien que le diagnostic des roulements par analyse vibratoire soit un domaine très développé, la recherche sur les représentations parcimonieuses des signaux de vibration est encore nouvelle et difficile pour le diagnostic des machines tournantes. Dans cette thèse, de méthodes nouvelles ont été développées, au moyen de différents modèles stochastiques, associées à des algorithmes efficaces afin de servir l’industrie dans le diagnostic des roulements. Tout d’abord, les modèles parcimonieux présentés dans la littérature sont revus. Les principales publications concernant le diagnostic des machines tournantes ont également été considérées. Enfin, en discutant des avantages et des inconvénients des représentations parcimonieuses, une interprétation des structures creuses d’un point de vue Bayésien est proposée, ce qui donne lieu à deux nouveaux modèles de diagnostic des machines tournantes. Dans un second temps, un nouveau modèle stochastique est proposé : il introduit une variable cachée relative à l’apparition d’impacts et estime le contenu spectral des transitoires correspondants ainsi que le spectre du bruit de fond. Cela donne lieu à un algorithme de détection automatique - sans besoin de pré-filtrage manuel - à partir duquel les fréquences de défaut peuvent être révélées. Le même algorithme permet également de filtrer le signal de défaut de manière très efficace par rapport à d’autres approches basées sur l’hypothèse stationnaire. La performance de l’algorithme est étudiée sur des signaux synthétiques. L’efficacité et la robustesse de la méthode sont également vérifiées sur les signaux de vibration mesurés sur un banc d’essai (engrenages et paliers). Les résultats sont meilleurs ou au moins équivalents à ceux de l’analyse d’enveloppes classique et du kurtogramme rapide. Dans un troisième temps, un nouveau schéma pour l’extraction de signaux cyclostationnaires (CS) est proposé. En considérant la variance périodique en tant que variable cachée, un filtre temporel est conçu de manière à obtenir la reconstruction intégrale des signaux CS caractérisés par une fréquence cyclique préétablie, qui peut être connue à priori ou estimée à partir de la corrélation spectrale. Un intérêt particulier de la méthode est sa robustesse lorsqu’elle est appliquée sur des données expérimentales ainsi qu’une capacité d’extraction supérieure par rapport au filtre de Wiener conventionnel. Finalement, ces exemples expérimentaux témoignent de l’utilisation polyvalente de la méthode à des fins de diagnostic de signaux composés. Pour finir, une analyse comparée utilisant le calcul rapide de la corrélation spectrale est réalisée sur une base de données publiquement disponible et largement utilisée. C’est un point crucial qui fixe un défis non-trivial à résoudre. / Although vibration-based rolling element bearing diagnostics is a very well-developed field, the research on sparse representations of vibration signals is yet new and challenging for machine diagnosis. In this thesis, several novel methods have been developed, by means of different stochastic models, associated with their effective algorithms so as to serve the industry in rolling element bearing diagnostics. First, the sparsity-based model (sparse code, in natural image processing) is investigated based on the current literature. The historical background of sparse representations has been inquired in the field of natural scenes. Along three aspects, its mathematical model with corresponding algorithms has been categorized and presented as a fundamental premise; the main publications are therefore surveyed in the literature on machinery fault diagnosis; finally, an interpretation of sparse structure in the Bayesian viewpoint is proposed which then gives rise to two novel models for machinery fault diagnosis. Second, a new stochastic model is introduced to address this issue: it introduces a hidden variable to indicate the occurrence of the impacts and estimates the spectral content of the corresponding transients together with the spectrum of background noise. This gives rise to an automatic detection algorithm – with no need of manual prefiltering as is the case with the envelope spectrum – from which fault frequencies can be revealed. The same algorithm also makes possible to filter out the fault signal in a very efficient way as compared to other approaches based on the stationary assumption. The performance is investigated on synthetic signals with a high noise-to-signal ratio and also in the case of a mixture of two independent transients. The effectiveness and robustness of the method are also verified on vibration signals measured on a test-bench (gears and bearings). Results are found superior or at least equivalent to those of conventional envelope analysis and fast kurtogram. Third, a novel scheme for extracting cyclostationary (CS) signals is proposed. By regularizing the periodic variance as hidden variables, a time-varying filter is designed so as to achieve the full-band reconstruction of CS signals characterized by some pre-set characteristic frequency. Of particular interest is the robustness on experimental data sets and superior extraction capability over the conventional Wiener filter. It not only deals with the bearing fault at an incipient stage, but it even works for the installation problem and the case of two sources, i.e. bearing and gear faults together. Eventually, these experimental examples evidence its versatile usage on diagnostic analysis of compound signals. Fourth, a benchmark analysis by using the fast computation of the spectral correlation is provided. One crucial point is to move forward the benchmark study of the CWRU data set by uncovering its own unique characteristics.

This diploma thesis deals with the design and creation of an algorithm for generating simulated signal data from a vibration diagnostics device. The first part is focused on theoretical acquaintance with vibration diagnostics and characteristics of individual defects of rotary machines. The next part deals with the possibilities of mathematical and kinematic simulations using a computer software. The main part of this work is dedicated to design and creation of software for generating simulated signal data. In the last part, the principle of simulation of specific defects of rotary machines is clearly demonstrated.

EDF exploite sur son parc de production de nombreuses machines électriques tournantes. Les contraintes thermiques subies par celles-ci engendrent des échauffements locaux qui nuisent à leur intégrité. Le présent travail contribue à fournir des méthodes de calcul adaptées à la détection et à la localisation des points chauds. Il participe à améliorer la compréhension des écoulements en rotation et leurs effets sur les transferts thermiques. Plusieurs dispositifs expérimentaux, de complexité ascendante, ont été utilisés pour comprendre et valider les simulations numériques. Une première étude sur une marche descendante (demi-pôle) parcourue par un jet de paroi non-confiné a mis en avant des différences par rapport à un jet confiné ; ces deux cas existent dans un alternateur. Une seconde étude menée sur une cavité tournante confinée a analysé l’impact d’un écoulement de Taylor-Couette-Poiseuille sur la température et la position des points chauds créés, en balayant l’ensemble des régimes d’écoulement. Ces études ont mis en exergue une première méthode de calcul fiable, fondée sur l’étude numérique CHT. Une autre méthode, basée sur la FEM couplée à une méthode inverse, a été testée sur une maquette d’alternateur hydraulique afin de pallier aux temps de calcul longs de la première. Cette méthodologie remonte aux coefficients d’échanges convectifs numériques à partir des mesures du champ thermique du rotor, mais n’est envisageable que lorsque l’on dispose de données expérimentales suffisantes. Ces travaux ont aussi mis en évidence de nouvelles techniques de mesures sans contact, comme l’utilisation d’un pyromètre à haute fréquence pour la mesure de température sur des machines tournantes. / EDF operates a large number of electrical rotating machines in its electricity generation capacity. Thermal stresses which affect them can cause local heating, sufficient to damage their integrity. The present work contributes to provide methodologies for detecting hot spots in these machines, better understanding the topology of rotating flows and identifying their effects on heat transfer. Several experimental scale model were used by increasing their complexity to understand and validate the numerical simulations. A first study on a turbulent wall jet over a non-confined backward-facing step (half-pole hydrogenerator) notes significant differences compared to results from confined case : both of them are present in an hydrogenerator. A second study was done on a small confined rotating scale model to determinate the effects of a Taylor-Couette-Poiseuille on temperature distribution and position of hot spots on the heated rotor, by studying the overall flow regimes flow. These studies have helped to obtain a reliable method based on conjugate heat transfer (CHT) simulations. Another method, based on FEM coupled with the use of an inverse method, has been studied on a large model of hydraulic generator so as to solve the computation time issue of the first methodology. It numerically calculates the convective heat transfer from temperature measurements, but depends on the availability of experimental data. This work has also developped new no-contact measurement techniques as the use of a high-frequency pyrometer which can be applied on rotating machines for monitoring temperature.

Ce travail de recherche s'inscrit dans le cadre de la surveillance des machines tournantes en régime non stationnaire et plus particulièrement la détection des défauts de roulement. Il se focalise sur la modélisation et l'analyse des variations de la vitesse de rotation instantanée. Les modèles numériques développés s'appuient sur une approche originale dite " angulaire " qui introduit explicitement les degrés de rotation libres de la machine et permet de s'affranchir de condition de fonctionnement en régime stationnaire. Un modèle de roulement à billes à gorge profonde a été développé couplant les efforts de contact normaux et tangentiels grâce à l'introduction du phénomène de résistance au roulement. Ce couplage permet d'expliquer l'origine des fluctuations de vitesse de rotation en présence de défaut de bague extérieure, phénomènes constatés expérimentalement : la présence d'un défaut sur une bague modifie périodiquement le couple de frottement. La modélisation de la liaison par engrenages est plus classique, la problématique a été de s'assurer qu'elle était compatible avec l'approche angulaire utilisée dans ces travaux. Le couplage des éléments de roulement et d'engrenage dans un modèle simple a montré que la présence de défauts de roulement provoque des variations des efforts radiaux ce qui modifie les efforts d'engrènement et en conséquence perturbe également la vitesse de rotation. Un modèle simplifié d'une éolienne a été réalisé et utilisé pour tester et valider des indicateurs spectraux de surveillance en conditions de fonctionnement non stationnaires. La comparaison avec des mesures expérimentales issues d'une éolienne a montré la robustesse du modèle. Il peut donc être exploité comme un outil d'analyse phénoménologique et de test d'indicateurs de surveillance. Il a par exemple permis de montrer l'importance des conditions de fonctionnement (charge et vitesse) sur le niveau des indicateurs utilisés en surveillance. Si besoin, les développements proposés peuvent être étendus sans difficulté à des modèles plus complexes de roulements, d'engrènements et de structures d'accueil de ces éléments technologiques. / This work has been developed within the framework of non-stationary rotating machinery surveillance with emphasis on the detection of roller bearing defects. It focuses on the modeling and analysis of instantaneous angular speed variations. The numerical models developed are based on an ``angular" approach which introduces explicitly the machine's free body rotation degrees of freedom allowing simulations in non stationary operating conditions. A deep groove ball bearing model has been developed coupling tangential and normal forces by taking into account the rolling resistance phenomenon. This coupling allows to explain the origin of angular speed fluctuations in the presence of outer race roller bearing defects, phenomena that has been observed experimentally: a roller bearing defect periodically modifies the friction torque. The description of the gear mesh has been performed by means of a classic approach where the scientific issue has been to verify the compatibility with the angular modeling, cornerstone of this dissertation. Integration of the roller bearing and the gear modeling approaches into a simple mechanical architecture has shown that the presence of bearing defects provoking variations in the rolling element-races normal forces, modifies the gear mesh forces leading to perturbations of the angular speed. A simplified wind turbine model has been used for the test and validation of non stationary surveillance spectral indicators. The robustness of the model has been shown by means of a comparison with measurements performed on a wind turbine. It has allowed to show the importance of the operating conditions (load and speed), on the level of indicators used. This means the model may be used as a tool for phenomenological analyses as well as a device to test surveillance indicators. If needed, the approaches presented may be easily extended into more complex models of bearings, gears, casings and foundations.

This master's thesis deals with order analysis. The first part of the thesis describes common methods for order analysis and methods for processing tacho and vibration signal. The second part contains a brief description of some open source tools for order analysis focusing on Sound and Vibration Measurement Suite (SVMS) package for LabVIEW by National Instruments company. The main purpose of the thesis has been designing and realization of own library for order analysis as a plugin for development tool LabVIEW. In the library, there are implemented methods for order analysis including basic functions for processing vibration and tacho signal. Their brief description with manual for its usage is in the third part of the thesis. In the last chapter, implemented functions are compared with those, which are implemented in SVMS by National Instruments, using both simulated and real data.

High Voltage (HV) rotating machines are widely used in power generation and heavy industry companies. Silicon Carbide (SiC) material is used in the stress grading structure of HV electric machines because of its non-linear characteristic. In this thesis, the resistance characteristics of SiC-based stress grading tape in time domain and the dielectric parameters in frequency domain were measured. The measurement results of samples with axial current-direction in time domain and frequency domain are about ten times that of the samples with radial current-direction, which might be a feature of the tape construction although it also can depend on the measurement method. The finite element method (FEM) simulation of the coil insulation system was carried out by using the experimental parameters, and the results were compared with the actual measurement results. The calculation results show that the non-linearity of SiC-based material does not have much effect on the real part of frequency response of the whole coil. The relative difference of the real part of complex capacitance is less than 5%. However, the influence on the imaginary part is relatively large, the relative difference is about 4% - 60%. / Högspänningsmaskiner används ofta i kraftproduktion och tunga industriella anläggningar. Kiselkarbidmaterial (SiC) används i sådana maskiners potential-styrningssstruktur på grund av dess icke-linjära egenskaper. I denna avhandling mättes ledningsförmågan hos SiC-baserade fältstyrningstejp i tidsdomänen och de dielektriska parametrarna i frekvensdomänen. Resultaten med axiell strömriktning vid tiddomän och frekvensdomänmetoder tyder på omkring tio gånger så stor konduktivitet jämfört med den radiella ström-riktningen, vilken kan bero på tejpkonstruktionen fast även eventuellt på mätmetoden. Simuleringen av en statorhärvas isoleringssystem utfördes med finita element metoden (FEM) med parametrar från materialproven, och resultaten jämfördes med mätresultat från en sådan härva. Beräkningsresultaten visar att det SiC-baserade materialets ickelinjäritet inte har någon större effekt på hela spolens verkliga frekvensrespons. Den relativa skillnaden mellan den verkliga delen av komplex kapacitans är mindre än 5%. Påverkan på den imaginära delen är dock relativt stor, med en relativ skillnad på ca 4% - 60%.

High voltage rotating machines play a significant role in generation and use of electrical energy as the demand for power continues to increase. However, one of the main causes for down times in high voltage rotating machines is related to problems with the winding insulation. The utilities want to reduce costs through longer maintenance intervals and a higher lifetime of the machines. These demands create a challenge for the producers of winding insulations, the manufacturers of high voltage rotating machines and the utilities to develop new insulation materials which can improve the life of the equipment and reduce the maintenance cost. The advent of nanotechnology in recent times has heralded a new era in materials technology by creating opportunities to significantly enhance the properties of existing conventional materials. Polymer nanocomposites belong to one such class of materials that exhibit unique combinations of physical, mechanical and thermal properties which are advantageous as compared to the traditional polymers or their composites. Even though they show tremendous promise for dielectric/electrical insulation applications, there are no studies relating to the long term performance as well as life estimation of the nanocomposites. Considering this, an attempt is made to generate an understanding on the feasibility of these nanocomposites for electrical insulation applications. An epoxy based nanocomposite system is chosen for this study along with alumina (Al2O3) and silica (SiO2) as the nanofillers. The first and the foremost requirement for studies on polymer nanocomposites is to achieve a uniform dispersion of nanoparticles in the polymer matrix, as nanoparticles are known to agglomerate and form large particle sizes. A laboratory based direct dispersion method is used to process epoxy nanocomposites in order to get well dispersed samples. A detailed microscopy analysis of the filler dispersion using Scanning Electron Microscope (SEM) has been carried out to check the dispersion of the nanofiller in the polymer. An attempt is made to characterize and analyze the interaction dynamics at the interface regions in the epoxy nanocomposite by glass transition temperature (Tg) measurements and Fourier transform infrared (FTIR) spectroscopy studies. The values of Tg for the nanocomposites studied decreases at 0.1 wt% filler loading and then starts to increase gradually with increase in filler loading. This Tg variation suggests that there is certainly an interaction between the epoxy chains and the nanoparticles. Also no new chemical bonds were observed in the spectra of epoxy nanocomposite as compared to unfilled epoxy. But changes were observed in the peak intensity and width of the –OH band in the spectrum of epoxy nanocomposite. This change was due to the formation of the hydrogen bonding between the epoxy and the nanofiller. The thermal conductivity of the epoxy alumina and the epoxy silica nanocomposites increased even with the addition of 0.1 wt% of the filler. This increase in thermal conductivity is one of the factors that make these nanocomposites a better option for electrical insulation applications. The dielectric properties of epoxy nanocomposites obtained in this investigation also reveal few interesting behaviors which are found to be unique and advantageous as compared to similar properties of unfilled materials. It is observed that the addition of fillers of certain loadings of nanoparticles to epoxy results in the nanocomposite permittivity value to be lower than that of the unfilled epoxy over the entire range of frequencies [10-2-106 Hz] considered in this study. This reduction has been attributed to the inhibition of polymer chain mobility caused by the addition of the nanoparticles. The tan values are almost the same or lower as compared to the unfilled epoxy for the different filler loadings considered. This behavior is probably due to the influence of the interface as the strong bonding at the interface will make the interface very stable with fewer defects apart from acting as charge trapping centres. From a practical application point of view, the surface discharge resistant characteristics of the materials are very important and this property has also been evaluated. The resistance to surface discharge is measured in the form of roughness on the surface of the material caused by the discharges. A significant enhancement in the discharge resistance has been observed for nanocomposites as compared to unfilled epoxy/ microcomposites, especially at longer exposure durations. The partial discharge (PD) measurements were carried out at regular intervals of time and it is observed that the PD magnitude reduced with discharge duration in the case of epoxy alumina nanocomposites. An attempt was made to understand the chemical changes on the surface by conducting the FTIR studies on the aged surface. For all electrical insulation applications, materials having higher values of dielectric strengths are always desired and necessary. So AC breakdown studies have also been conducted. The AC breakdown strength shows a decreasing trend up to a certain filler loading and then an increase at 5 wt% filler loading for epoxy alumina nanocomposites. It has been also observed that the type of filler as well as the thickness of the filler influences the breakdown strength. The AC dielectric strength of microcomposites are observed to be lower than the nanocomposites. Extensive research by long term aging studies and life estimation are needed before these new nanocomposites can be put into useful service. So long term aging studies under combined electrical and thermal stresses have been carried out on unfilled epoxy and epoxy alumina nanocomposite samples of filler loading 5 wt%. The important dielectric parameters like pemittivity, tan  and volume resistivity were measured before and after aging to understand the performance of the material under study. The leakage current was measured at regular intervals and tan  values were calculated with duration of aging. It was observed that the tan  values increased drastically for unfilled epoxy for the aging duration considered as compared to epoxy alumina nanocomposites. The life estimation of unfilled epoxy as well as epoxy nanocomposites were also performed by subjecting the samples to different stress levels of 6 kV/mm, 7 kV/mm and 8 kV/mm at 60 oC. It is observed that the epoxy alumina nanocomposite has an enhanced life which is nine times the life of the unfilled epoxy. These results obtained for the nanocomposites enable us to design a better material with improved dielectric strength, dielectric properties, thermal conductivity, resistance to surface discharge degradation and enhanced life without sacrificing the flexibility in the end product and the ease of processing. Dry type transformers and stator winding insulation need to be cast with the above material developed and tested before practically implementing these in the actual application.