Electric propulsion motors have emerged as a promising solution to address greenhouse gas emissions from Internal Combustion Engines (ICEs). While electric propulsion motors offer numerous advantages over Internal Combustion Engines (ICEs), they also pose certain challenges. Electric motors are prone to high-frequency tonal noise, which can be annoying to customers and become a quality concern in noise-sensitive automotive applications. The ongoing effort to increase the speed of electric propulsion motors for enhanced power density can have an adverse impact on rotors. This is due to the fact that the stress induced in the rotor is quadratically proportional to its speed. This concern becomes particularly significant for motors that rely on air barriers and thin bridges to enhance their electromagnetic performance. The thesis makes a contribution to address these challenges. First, the acoustic noise mitigation methods at the transmission stage are investigated. Then, acoustical materials are experimentally validated for their capacity to mitigate acoustic noise at the transmission stage. Then, experimental modal analysis is conducted to find out the modal characteristics of a stator-housing assembly. The mode shapes and modal frequency are compared with finite element results to evaluate the fidelity of the finite element model. Then, an equivalent damage approach is used to employ accelerated fatigue analysis for a rotor using constant amplitude load cycles. Finally, a thermomechanical fatigue analysis workflow is developed for a rotor to overcome the limitations of the constant amplitude load cycle approach, with an additional computational cost. / Thesis / Doctor of Philosophy (PhD)

The United States Air Force is in the process of developing a more electric aircraft. The development of an aircraft Integrated Power Unit and an Internal Starter/Generator will be instrumental in producing sufficient electrical power to run all non propulsive systems. Iron-cobalt alloys, such as Hiperco alloy 50HS, are high temperature, high strength magnetic materials ideal for these power applications. Design requirements and previous studies indicate that these materials need to survive in temperatures up to 1000F (810K), rotation speeds of about 55,000 rpm, and have strengths in excess of 80 ksi. Research conducted by Fingers provided the material and creep properties used in the analysis presented in this report. The finite element method was used to analyze a spinning rotor mounted to a circular shaft via an interference fit subjected to various operating environments. The power law creep model defined by Fingers was used to analyze three distinct rotor configurations. The first configuration was a constant temperature single lamina, mounted to a shaft of equal thickness, subject to temperatures between 727K and 780K, rotation speeds between 35,000 rpm and 60,000 rpm, and two different interference fits: 0.0015 inches and 0.003 inches. The results yield conservative predictions that indicate that these models could not survive the required operating conditions. The second configuration was a linear radial variation in temperature single lamina, mounted to a shaft of equal thickness, subjected to three temperature ranges, rotation speeds between 30,000 rpm and 55,000 rpm, and two different interference fits; 0.0015 inches and 0.003 inches. These results represent a more realistic model, which indicate that the "cooler" inner portions of the rotor restrict the creep deformations of the "hotter" outer portions resulting in higher possible operating temperatures and rotation speeds very near the required operating conditions. The third configuration was a lamina stack comprised of two rotor lamina, with a Coulomb friction surface interaction, and held together by a compressive axial force. These models represent a first step towards understanding the behavior of the entire rotor stack. / Master of Science

A stator cascade was developed to simulate the flow conditions within a close-stage-spacing transonic axial compressor. Experiments were conducted in a linear transonic blowdown cascade wind tunnel with an inlet Mach number of 0.65. The bow shock from the downstream rotor was simulated by a single moving normal shock generated with a shock tube. First, steady pressure data were gathered to ensure that the stator cascade operated properly without the presence of the shock. Next, the effects of the passing shock on the stator flow field were investigated using shadowgraph photography and Digital Particle Image Velocimetry (DPIV). Measurements were taken for three different shock strengths. In every case studied, a vortex formed near the stator trailing edge as the shock impacted the blade. The size of this vortex was shown to be directly related to the shock strength, and the vortex remained present in the trailing edge flow field throughout the cycle duration. Analysis of the DPIV data showed that the vortex acts as a flow blockage, with the extent of this blockage ranging from 2.9% of the passage for the weakest shock, to 14.3% of the passage for the strongest shock. The vortex was also shown to cause flow deviation up to 75° for the case with the strongest shock. Further analysis estimated that the total pressure losses due to shock-induced vorticity ranged from 46% to 113% of the steady wake losses. Finally, the total pressure loss purely due to the upstream-propagating normal shock was estimated to be roughly 0.22%. / Master of Science

Según diversos estudios, los fallos más comunes en motores de inducción se dan en rodamientos, en el aislamiento del devanado estatórico y, en menor medida, en el rotor. Este último tipo de fallo incluye la rotura, agrietamiento o deformación de las barras o de los anillos de cortocircuito de la jaula rotórica. A pesar de que el porcentaje de fallos relacionados con el rotor pueda parecer no demasiado elevado (algunos autores lo sitúan en torno al 15-20% del total de averías), se han reportado casos en los que la avería sí ha conducido a fallos catastróficos inmediatos, como situaciones de protrusión de barras o de desprendimiento de fragmentos de barras rotas que han acabado dañando el aislamiento del estator. Estas averías han provocado pérdidas millonarias para las compañías involucradas. Finalmente, este tipo de fallo es más frecuente en motores de gran potencia. Son estos motores los más caros, los más costosos de reparar y, usualmente, los más críticos en las aplicaciones en las que operan. Uno de los retos que se ha abordado en la presente tesis consiste en la aplicación y validación generalizada de la metodología basada en el análisis de la corriente de arranque para el caso de motores arrancados mediante arrancadores estáticos. A este respecto, se ha constatado el uso creciente de arrancadores estáticos en multitud de aplicaciones industriales. La presente tesis aborda por primera vez esta problemática de forma rigurosa. La tesis se centra en la validación de una variante concreta de diagnóstico transitorio, la cual se basa en el análisis de la señal de corriente de arranque mediante la transformada wavelet discreta (Discrete Wavelet Transform, DWT). Se trata ésta de una transformada tiempo-frecuencia que cuenta con importantes ventajas a la hora de emplear la metodología de diagnóstico transitorio, dado que esta transformada resulta rápida de aplicar, la interpretación de sus resultados es simple y está disponible en paquetes comerciales convencionales como Matlab, con lo que cualquier usuario con un mínimo entrenamiento podría hacer uso de ella. Adicionalmente, esta transformada permite la introducción sencilla de indicadores para determinar el grado de severidad de la posible avería en el rotor, cuestión ésta que no resulta trivial con otras alternativas. Se han obtenido más de 900 señales de corriente de arranque en motores de laboratorio, utilizando hasta cinco modelos diferentes de arrancadores de distintos fabricantes y con variadas topologías y características. Para cada arrancador se han obtenido múltiples señales correspondientes a diferentes niveles de fallo en el rotor (motor sano, motor con una barra rota y motor con dos barras rotas). Para cada nivel de fallo, se han obtenido una variedad de señales mediante la variación de los parámetros de arranque (tensión inicial, duración de rampa), así como de otras características operativas del motor (nivel de carga, tensión de suministro). Adicionalmente, se han obtenido múltiples señales en motores industriales accionados mediante arrancador, que operaban en industrias (alimentación, plantas de depuración), con el fin de ratificar la validez de los resultados obtenidos con los de laboratorio. Asimismo, se ha incluido un epígrafe final que ahonda en la automatización a la hora de aplicar la metodología propuesta, de forma que se evite, en la medida de lo posible, la necesidad de intervención de usuarios expertos para su aplicación. Este epígrafe se ha basado en una investigación conjunta con otros grupos internacionales. Los resultados son prometedores, por cuanto prueban la validez de la metodología para detectar, no solamente la presencia de la avería, sino también para determinar la severidad del fallo en el rotor. Especialmente interesante es el hecho de que los resultados son más concluyentes si cabe en el caso de motores con arrancadores con control en las tres fases, que no / According to various studies, the most common failures in induction motors are bearing failures, failures of the isolation of the winding stator and, to a lesser extent, rotor failure. The latter failure includes the breaking, cracking or deformation of the bars or the rings of the shortcircuit of the rotor cage. Although the percentage of failures related to the rotor may not seem high (some authors place it around 15-20% of breakdowns), there have been reported cases in which the breakdown has resulted in immediate catastrophe, situations in which the protrusion of the bars or the detachment of fragments of broken bars that have ended up damaging the isolation of the stator. These breakdowns have caused the loss of millions of euros for the companies involved. Finally, this type of failure is more frequent in motors with high horsepower. It is these motors, the most costly to repair, and usually the most critical to the applications in which they operate. One of the challenges that has been addressed in this thesis consists of the application and generalized validation of the methodology based on the analysis of the starting current for the case of engines started by static starters (also called electronic starters or soft starters). In this regard, the increasing use of static starters in a multitude of industrial applications has been verified. This thesis deals with this problem in a rigorous way for the first time. The thesis focuses on the validation of a specific variant of transient diagnosis, which is based on the analysis of the start current signal by the discrete transformed wavelet (DWT). It is a time-frequency transformation that has important advantages when using the methodology of transient diagnosis. Given that this transformation is quick to apply, the interpretation of its results is simple and available in conventional commercial packages like Matlab, which any user with a minimum training could utilize. Additionally, this transformer allows the simple introduction of indicators to determine the degree of severity of the possible breakdown in the rotor, a question that is not trivial with other alternatives. More than 900 signals have been obtained of starting current in laboratory motors, using up to five models of different starters from different manufacturers and with varied topologies and characteristics. Multiple corresponding signals have been obtained for each starter at different levels of rotor failure (healthy motor, motor with a broken bar and motor with two broken bars). For each level of failure, a variety of signals have been obtained by varying the starting parameters (initial tension, duration of ramp ...), as well as other operating characteristics of the motor (load level, voltage of supply). Additionally, multiple signals have been obtained in industrial engines powered by a starter, which operated in industries of very diverse sectors (food, depuration plants), in order to ratify the validity of the results obtained with the laboratory engines. Likewise, a final section has been included that delves into automation when applying the proposed methodology, so that, as far as possible, the need for expert user intervention for their application can be avoided. This section has been based on joint research with other international groups. The results are promising, as they prove the validity of the methodology to detect, not only the presence of the breakdown, but also to determine the severity of the failure in the rotor. Especially interesting is the fact that the results are more conclusive in the case of engines with starters with control in all three phases, which are what are normally used for more powerful motors. / Segons diversos estudis, les fallades més comunes en motors d'inducció són les fallades en rodaments, les fallades en l'aïllament del debanament estatòric i, en menor mesura, les fallades en el rotor. Aquest últim tipus de fallada inclou el trencament, clavillament o deformació de les barres o dels anells de curtcircuit de la gàbia rotòrica. Tot i que el percentatge de fallades relacionades amb el rotor puga semblar no massa elevat (alguns autors ho situen al voltant del 15-20% del total d'avaries), s'han reportat casos en els quals l'avaria sí ha conduït a fallades catastròfiques immediates, com a situacions de protrusió de barres o de despreniment de fragments de barres trencades que han acabat danyant l'aïllament de l'estator. Aquestes avaries han provocat pèrdues milionàries per a les companyies involucrades. Finalment, aquest tipus de fallada és més freqüent en motors de gran potència. Són aquests motors els més cars, els més costosos de reparar i, usualment, els més crítics en les aplicacions en les quals operen. Un dels reptes que s'ha abordat en la present tesi consisteix en l'aplicació i validació generalitzada de la metodologia basada en l'anàlisi del corrent d'arrancada per al cas de motors arrancats mitjançant arrancadors estàtics. Referent a açò, s'ha constatat l'ús creixent de arrancadors estàtics en multitud d'aplicacions industrials. La present tesi aborda per primera vegada aquesta problemàtica de forma rigorosa. La tesi se centra en la validació d'una variant concreta de diagnòstic transitori, la qual es basa en l'anàlisi del senyal de corrent d'arrancada mitjançant la transformada wavelet discreta (Discrete Wavelet Transform, DWT). Es tracta aquesta d'una transformada temps-freqüència que compta amb importants avantatges a l'hora d'emprar la metodologia de diagnòstic transitori, atès que aquesta transformada resulta ràpida d'aplicar, la interpretació dels seus resultats és simple i està disponible en paquets comercials convencionals com Matlab, amb el què qualsevol usuari amb un mínim entrenament podria fer-ne ús. Addicionalment, aquesta transformada permet la introducció senzilla d'indicadors per a determinar el grau de severitat de la possible avaria en el rotor, qüestió aquesta que no resulta trivial amb altres alternatives. S'han obtingut més de 900 senyals de corrent d'arrancada en motors de laboratori, utilitzant fins a cinc models diferents de arrancadors de diferents fabricants i amb variades topologies i característiques. Per a cada arrancador s'han obtingut múltiples senyals corresponents a diferents nivells de fallada en el rotor (motor sa, motor amb una barra trencada i motor amb dues barres trencades). Per a cada nivell de fallada, s'han obtingut una varietat de senyals mitjançant la variació dels paràmetres d'arrancada (tensió inicial, durada de rampa), així com d'altres característiques operatives del motor (nivell de càrrega, tensió de subministrament). Addicionalment, s'han obtingut múltiples senyals en motors industrials accionats mitjançant arrancador, que operaven en indústries (alimentació, plantes de depuració), amb la finalitat de ratificar la validesa dels resultats obtinguts amb els motors de laboratori. Així mateix, s'ha inclòs un epígraf final que aprofundeix en l'automatització a l'hora d'aplicar la metodologia proposada, de manera que s'evite, en la mesura del possible, la necessitat d'intervenció d'usuaris experts per a la seua aplicació. Aquest epígraf s'ha basat en una recerca conjunta amb altres grups internacionals. Alhora, els resultats són prometedors, puix que proven la validesa de la metodologia per a detectar, no sols la presència de l'avaria, sinó també per a determinar la severitat de la fallada en el rotor. Especialment interessant és el fet que els resultats són més concloents, si açó és possible, en el cas de motors amb arrancadors amb control en les tres fases, que són / Corral Hernández, JÁ. (2018). Aportaciones al diagnóstico de averías en motores eléctricos basados en el análisis avanzado de corrientes [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/106368

Due to the ever-increasing penetration of distributed generation units connected to the power distribution system, electric power systems, worldwide, are undergoing a paradigm shift with regards to system monitoring, operation and control. We envision that with the emergence of `active' distribution systems consisting of `prosumers' and localized energy markets, decentralized control methods in power systems are gaining a growing attention among power researchers. Traditionally, two main types of control schemes have been implemented in power systems: (a) wide-area monitoring based centralized control, and (b) local measurement based primary (machine) level control. By contrast, decentralized control schemes based on local monitoring and control of strategically-determined subsystems (or `areas') of a large-scale power system are not used. The latter control schemes offer several advantages over the former, which include more flexibility, simplicity, economy and scalability for large-scale systems. In this dissertation, we summarize our research work on hierarchical and decentralized control techniques for the enhancement in a unified manner of voltage and rotor angle stability in large-scale power systems subject to large (e.g., short circuits) and small (e.g., small load changes) disturbances. We study system robustness by calculating local stability margins. We derive decentralized control laws that guaranty global asymptotic stability by applying Lyapunov's second method for interconnected systems. Furthermore, we argue that the current centralized control structure must only play a supervisory control role at a higher (tertiary) hierarchical level by processing the decisions taken by the regional control entities regarding the stability/instability of the system. This ensures system-wide situational awareness while minimizing the communication bandwidth requirements. We also develop a multi-agent based framework for this hierarchical control scheme. Finally, we compare different communication protocols using simulation models and propose an efficient communication network design for decentralized control schemes. This work, in principle, motivates the development of fast stability analysis which, in the future, may also account for the non-linear coupling that exist between machine rotor angles and bus voltages in power system models. As a future work, we propose the use of statistical techniques like random-effects regression and saddlepoint approximation method to reliably estimate the type-I and type-II probability errors in the proposed hierarchical, decentralized control decision process. / Ph. D. / In the present research work, we have proposed a decentralized, hierarchical control scheme for large-scale, interconnected power systems. Using Lyapunov’s second method for interconnected systems, we have derived <i>decentralized</i> control laws for control devices which ensure global asymptotic stability of weakly interconnected power systems. The decentralized control schemes have several advantages over centralized ones. For instance, the former approaches lead to a reduction of dimensionality in terms of both modeling and controlling the state variables of a system. One of the major contributions of this research work is the reduction of the dimensionality of the energy function and of the regression models used to determine the control laws, which results from the non requirement of exchanging information between control areas or of sending all the measurements to a central controller. Hence, the proposed scheme also yields smaller communication requirements than those of completely centralized wide-area controllers, while offering a better situational awareness. Similar decentralized control schemes are commonly used in robotic, transportation and surveillance systems.

The effect of flow anisotropy and three dimensional separation on the turbulent structure and radiated turbulence ingestion noise of a rotor in two experimental configurations is studied. The first consists of a non-axisymmetric boundary layer wake ingested by a rotor mounted at the aft of a body of revolution inclined at 5 degree angle of attack. In the second configuration a transient disturbance is generated by an upstream wing body junction pitching from zero to 20 degree angle of attack . This disturbance is convected downstream and ingested into a rotor immersed in a wall boundary layer. In both cases flow velocimetry at the rotor inflow is done and the far field sound is measured. The flow velocimetry in the wake of the inclined body of revolution shows evidence of three dimensional separation and vortex rollup between the lee and body sides. A boundary layer embedded shear layer develops as the turbulent kinetic energy is pulled off the wall by the flow separation and is visible in the port side velocimetry. The turbulent structure of this shear layer and the boundary layer on the lee of the body is visualized using compact eddy structure representation and the modes on the port side are shown to be stretched versions of similar modes seen in an equilibrium, zero pressure gradient boundary layer. The effect these structures had on the radiated sound served to both increase blade to blade correlation and the overall broadband levels of the sound. Measurements of the sound using an acoustic array showed directivity effects that resulted from the location of the embedded shear layer and rollup vortices. Although the vortices likely have some effect on the spectra, most of the noise is dominated by the turbulence ingestion of the embedded shear layer. For the second experimental configuration the transient motion was documented through repeated measurements of the flow field and sound, and an ensemble average of the measurements taken. Overall the flow was unsteady, particularly in the outer region of the boundary layer. The sound radiated was shown to be tonal during the first half of the interaction, where the flow is dominated by a deterministic mean flow change, and attributed to a form of periodic unsteady loading. During the latter half of the disturbance the broadband and overall sound levels increased significantly and are associated with the interaction of the rotor with flow separation over the wing body junction when it reached a critical, 16 degree angle of attack. / Doctor of Philosophy / The interaction of rotors and propellers with turbulence is commonly encountered when vehicles transit fluid mediums. In vehicles with aft mounted propellers, such as pusher type aircraft or underwater vehicles, turbulent boundary layers developed over the vehicle surface are ingested by the propeller. The size or scale of the average turbulent eddy greatly affects the type of sound generated by the interaction. For eddies that are small enough to only interact with one blade, the blade angle of attack varies randomly as it rotates through the turbulence and this radiates broadband sound. However, if the blade encounters eddies that are large or long enough to interact with multiple blades then the angle of attack, and thus the fluctuating lift force, begins to be correlated for each passage of the blade. This is known as blade to blade correlation and produces energy and sound concentration around the frequencies that correspond to the blade passage. This phenomenon is fairly well understood and many attempts have been made to model and predict the sound spectra from a rotor encountering turbulence in this manner. However these models often assume isotropic and homogenous turbulence when making predictions. This assumption works well in many applications, however, often the turbulence the rotor encounters is anisotropic with significant flow inhomogeneities. Thus, experimental investigations into the mechanisms and sources of sound in inhomogenous and anisotropic flows is necessary in an attempt to inform further flow and acoustic models. In this dissertation the inflow and acoustic response of a rotor ingesting significantly complex and anisotropic flows is characterized. It focuses on two commonly encountered flow arrangements; a rotor mounted at the stern of a body of revolution at an angle of attack, and a rotor ingesting a turbulent wall boundary layer with transient disturbances introduced by an upstream wing body junction. In both cases the flow is three dimensional and the rotor encounters significant circumferential turbulence variation during its rotation through the resultant turbulent flow field. For the flow about the body of revolution the flow and noise appear to be driven by the rotor interaction with an embedded shear layer that results from three dimensional separation between the lee and windward sides of the body. For the transient disturbance interaction the rotor noise response shows two separate noise sources. During the first half of the disturbance the blade response is tonal and associated with a deterministic blade angle of attack change as the rotor interacts with the transient. In the latter half of the disturbance the rotor broadband noise is significantly increased due to flow separation over the wing body junction.

This thesis focuses on generating real world turbulence levels in a water tunnel rotor test using fractal grids and characterizing the effect of the fractal grid generated-turbulence on the performance of hydrokinetic turbines. The research of this thesis is divided into three studies: one field study and two laboratory studies. The field study was conducted at the Canadian Hydro Kinetic Turbine Test Centre (CHTTC) on the Winnipeg River. An Acoustic Doppler Velocimeter (ADV) was used in the field study to collect flow measurements in the river. The laboratory studies were conducted at the University of Victoria (UVic) fluids research lab and the Sustainable Systems Design Lab (SSDL). In addition, the Particle Image Velocimetry (PIV) technique was used in the experiential studies to obtain quantitative information about the vector flow field along the test section, both upstream and downstream of the rotor’s plane. The first study is a field study aiming to provide real flow characteristics and turbulence properties at different depths from the free-surface to boundary layer region of a fast river current by conducting a field study in the Winnipeg River using ADV. A novel technique to deploy and control an ADV from free-surface to boundary layer in a fast-current channel is introduced in this work. Flow characteristics in the river, including mean flow velocities and turbulence intensity profiles are analyzed. The obtained results indicate that the maximum mean velocity occurs below the free-surface, suggesting that the mean velocity is independent of the channel depth. From the free-surface to half depth, it was found that changes in both the mean velocity and turbulence intensity are gradual. From mid-depth to the river bed, the mean velocity drops rapidly while the turbulence intensity increases at a fast rate. The turbulent intensity varied from 9% at the free-surface to around 17.5% near the river bed. The results of this study were used in the second lab study to help designing a fractal grid for a recirculating water flume tank. The goal was to modify the turbulence intensity in the water tunnel such that the generated turbulence was similar to that in the river at a location typical of a hydrokinetic device. The properties of fractal-generated turbulence were experimentally investigated by means of 2D Particle Image Velocimetry (PIV). The streamwise turbulent intensity profiles for different grids along the channel are presented. Additionally, visualization of the average and fluctuating flow fields are also presented. The results are in good agreement with results in literature. The third and final study investigated the power coefficient of a scale hydrokinetic turbine rotor in controlled turbulent flow (7.4 % TI), as well as in the low-turbulence smooth flow (0.5% TI) typical of lab scale testing. PIV was employed for capturing the velocity field. The results show that using realistic TI levels in the water tunnel significantly decrease the turbine’s power coefficient compared to smooth flow, highlighting the importance of considering this effect in future experimental campaigns. / Graduate

Le bruit tonal d’interaction rotor-stator, généré par l’impact des sillages issus des pales d’un rotorsur la grille d’aubes d’un stator redresseur, contribue de manière déterminante au bruit d’origineaérodynamique des turbomachines axiales carénées, qui équipent une large part des systèmes de propulsionaéronautique et de conditionnement d’air. La prédiction du bruit par l’utilisation de simulationsnumériques demeure onéreuse, notamment dans les premières phases de conception lorsque de nombreusesconfigurations doivent être testées. Dans cette optique, l’approche analytique choisie dans cettethèse apporte une alternative tout à fait appropriée. Les modèles analytiques basés sur une fonctionde réponse aéroacoustique de profil isolé ne permettent pas de reproduire l’effet de grille engendrépar le nombre important d’aubes de stator. Inversement, de fortes approximations sont nécessairespour décliner les fonctions de réponse de grilles d’aubes existantes dans des configurations tridimensionnelles.Le formalisme proposé, basé sur la méthode de raccordement modal, permet d’introduiresimplement l’effet de grille dans une géométrie annulaire d’étage rotor-stator. Un modèle de réponse degrille rectilinéaire bidimensionnel est tout d’abord présenté pour la transmission d’ondes acoustiques àtravers le stator ainsi que pour la génération de bruit par l’impact de rafales hydrodynamiques. Dansce cadre, une analyse linéaire et non visqueuse est considérée, pour laquelle les modes acoustique ettourbillonnaire d’un gaz sont couplés par le biais des frontières rigides. Les perturbations de vitessesliées aux sillages sont modélisées comme des rafales convectées. Leur impact sur la grille de statorgénère des ondes acoustiques se propageant en amont, en aval ainsi que dans les espaces inter-aubesdu stator, vu comme un réseau périodique de guides d’ondes. Les sections de bords d’attaque et defuite des aubes sont considérées comme des interfaces sur lesquelles la continuité des fluctuations depression, de vitesse axiale et de vorticité est vérifiée. Un système d’équations est ainsi obtenu, puisrésolu par des projections sur les bases modales du conduit et des inversions matricielles. Le champacoustique rayonné est ainsi déterminé uniformément dans tout le domaine. Les résultats issus de cesmodélisations sont comparés à ceux des fonctions de réponse de grilles d’aubes rectilinéaires issues dela littérature, montrant un très bon accord avec les modèles basés sur la technique de Wiener-Hopf. Leformalisme est par la suite étendu aux grilles annulaires par l’ajout de fonctions de Bessel comme fonctionsde forme radiale exprimant les effets tridimensionnels. Finalement, une procédure est présentéepour rendre compte de l’hétérogénéité des aubes de stator, caractéristique des nouvelles architecturesde turbomachines. Cette méthodologie est basée sur l’emploi conjoint du principe du dipôle de bordd’attaque et de la fonction de réponse aéracoustique de la grille de stator à l’aide de la technique deraccordement modal. Le principe de dipôle de bord identifie le chargement instationnaire des aubesinduit par l’impact de rafales hydrodynamiques, calculé par le formalisme d’Amiet, avec la trace duchamp de pression acoustique produit par un dipôle placé au voisinage du bord de l’aube. Les prédictionsissues de ce modèle, appliqué dans un cadre bidimensionnel, sont ensuite comparées à des mesuresobtenues pendant la campagne d’essais du projet SEMAFOR. / The rotor-stator wake-interaction tonal noise, generated by the impingement of rotor wakes onoutlet guide vanes, plays a crucial role in the aerodynamic noise of axial-flow ducted fan stages. Thelatter are widely used in most aeronautic propulsion and air-conditioning systems. The noise predictionby means of numerical simulations remains expensive, especially at the preliminary design stage whennumerous configurations must be tested. In this respect, the analytical approach chosen in this thesisprovides a well suited alternative. The analytical modeling based on an isolated-airfoil response functioncan not reproduce the cascade effect introduced by the large number of stator vanes. Conversely, drasticapproximations are required to extend the current cascade response functions to three-dimensionalconfigurations. The proposed modeling based on the mode-matching technique simply introduces thecascade effect in an annular rotor-stator stage. A rectilinear cascade response function is firstly presentedto account for the acoustic transmission through the stator along with the wake-interaction noise.In this context, a linearized and non-viscous analysis is carried out, in which the acoustic and vorticalmodes of a gas are coupled at rigid physical boundaries. The velocity perturbations issued from thewakes are written as a sum of convected gusts. Their impingement on the cascade of vanes generatesacoustic waves propagating upstream, downstream of the cascade, as well as inside the inter-vane channelsof the stator, seen as a periodic array of bifurcated waveguides. The duct cross sections at theleading-edge and the trailing-edge of the vanes are seen as interfaces on which the continuity of thefluctuating pressure, axial velocity and vorticity is fulfilled. A system of linear equations is obtained,then solved by means of modal projections and matrix inversions. The acoustic field is then uniformlycalculated in the whole domain. Comparisons with rectilinear cascade response functions show a verygood agreement with predictions based on the Wiener-Hopf technique. The configuration of an annularcascade is addressed by introducing the Bessel functions as radial shape functions, expressing threedimensionaleffects. Finally, a procedure is presented to account for the heterogeneity of the statorvanes, typical of modern fan architectures. This approach is based on the combinaison of the leadingedgedipole principle and the cascade response function derived from the mode-matching technique.The edge-dipole principle identifies Amiet’s solution for the unsteady loading and the radiation of adipole approached very close to the edge of a half plane. The predictions provided by this modeling,applied in a two-dimensional configuration, are finaly compared to measurements performed in the testcampaign of the SEMAFOR project.

Os motores de indução desempenham um importante papel na indústria, fato este que destaca a importância do correto diagnóstico e classificação de falhas ainda em fase inicial de sua evolução, possibilitando aumento na produtividade e, principalmente, eliminando graves danos aos processos e às máquinas. Assim, a proposta desta tese consiste em apresentar um multiclassificador inteligente para o diagnóstico de motor sem defeitos, falhas de curto-circuito nos enrolamentos do estator, falhas de rotor e falhas de rolamentos em motores de indução trifásicos acionados por diferentes modelos de inversores de frequência por meio da análise das amplitudes dos sinais de corrente de estator no domínio do tempo. Para avaliar a precisão de classificação frente aos diversos níveis de severidade das falhas, foram comparados os desempenhos de quatro técnicas distintas de aprendizado de máquina; a saber: (i) Rede Fuzzy Artmap, (ii) Rede Perceptron Multicamadas, (iii) Máquina de Vetores de Suporte e (iv) k-Vizinhos-Próximos. Resultados experimentais obtidos a partir de 13.574 ensaios experimentais são apresentados para validar o estudo considerando uma ampla faixa de frequências de operação, bem como regimes de conjugado de carga em 5 motores diferentes. / Induction motors play an important role in the industry, a fact that highlights the importance of correct diagnosis and classification of faults on these machines still in early stages of their evolution, allowing increase in productivity and mainly, eliminating major damage to the processes and machines. Thus, the purpose of this thesis is to present an intelligent multi-classifier for the diagnoses of healthy motor, short-circuit faults in the stator windings, rotor broken bars and bearing faults in induction motors operating with different models of frequency inverters by analyzing the amplitude of the stator current signal in the time domain. To assess the classification accuracy across the various levels of faults severity, the performances of four different learning machine techniques were compared; namely: (i) Fuzzy ARTMAP network, (ii) Multilayer Perceptron Network, (iii) Support Vector Machine and (iv) k-Nearest-Neighbor. Experimental results obtained from 13.574 experimental tests are presented to validate the study considering a wide range of operating frequencies and also load conditions using 5 different motors.

Orientador: Dr. Alfeu J. Sguarezi Filho / Dissertação (mestrado) - Universidade Federal do ABC, Programa de Pós-Graduação em Engenharia Elétrica, 2014. / Devido as preocupações com a redu¸c¿ao do uso dos recursos f'osseis e das emiss¿oes de CO2 ocasionadas pelos mesmos, o interesse no uso de energias renov 'aveis aumentou, sendo uma destas fontes a energia e'olica. O Gerador de Indu¸c¿ao de Rotor Bobinado (GIRB) tem demonstrado ser uma ferramenta capaz de otimizar o uso da energia e'olica, dada a sua alta efici¿encia, rentabilidade e robustez mec¿anica, ainda assim 'e necess'ario superar os numerosos desafios que est¿ao presentes nesta tecnologia, tais como qualidade da pot¿encia, estabilidade da rede, entre outros. O objetivo desta pesquisa se concentra no estudo de t'ecnicas de controle de pot¿encia fornecida pelo GIRB aplicado a sistemas de gera¸c¿ao e'olica. No presente caso, o sistema de gera¸c¿ao 'e composto por um GIRB com seu estator conectado diretamente 'a rede e seu rotor 'e conectado 'a rede atrav'es de um conversor bidirecional. Prop¿oe-se estudar o m'etodo de controle direto de torque (CDT) com emprego de controladores PI e por modos deslizantes (SMC) mais PI. O controle direto de torque 'e uma t'ecnica de alto desempenho din¿amico e possibilita o controle independente do torque e fluxo do gerador o que possibilitar'a o controle das pot¿encias ativa e reativa do GIRB. Da mesma forma, 'e estudado o conversor para o processamento da energia gerada. As simula¸c¿oes baseadas em modelos matem'aticos destes m'etodos de controle foram revistos, analisados e comparados com as simula¸c¿oes de textos de refer¿encia. Resultados experimentais obtidos em uma bancada validaram o prot'otipo proposto. / Due to concerns about reducing the use of fossil resources and the CO2 emissions caused by them, interest in the use of renewable energy increased, one of these sources is wind energy. The Wound Rotor Induction Generator (WRIG) has been proved to be a tool to optimize the use of wind energy, given its high efficiency, profitability and mechanical robustness, it is still necessary to overcome the many challenges that are present in this technology, such as power quality, grid stability, among others. This research focuses on the study of control techniques for power supply by WRIG applied to wind generation systems. In this case, the generation system consists of a stator WRIG is connected directly to the electrical network and its rotor too, but via a bidirectional converter. It is proposed to study the method of direct torque control (DTC) with the use of PI and sliding mode controllers (SMC) more PI. The direct torque control is a technique of high dynamic performance and enables independent control of torque and flux generator, which will enable the control of active and reactive power ofWRIG. Likewise, it is studied the converter for processing the generated power. The simulations based on mathematical models of these control methods were reviewed , analyzed and compared with simulations of reference texts . Experimental results obtained on a bench validated the proposed prototype.