Depuis une quinzaine d'années, l'éolien s'est grandement développé en nombre d'infrastructures et en puissance unitaire mais il reste toujours confronté à un problème de disponibilité de par les nombreuses pannes d'ordre mécanique ou électrique. Le but de ces travaux consiste à concevoir, modéliser et piloter des aérogénérateurs tolérants aux défauts mécaniques et électriques. Pour cela, une structure basée sur une génératrice asynchrone hexaphasée à grand nombre de paires de pôles a été retenue. L'augmentation du nombre de pôles permet de s'affranchir ou de simplifier le multiplicateur, source des pannes mécaniques, tandis que l'utilisation d'une structure multiphasée permet de poursuivre la production d'énergie lors de la perte de phases au stator ou de bras du convertisseur. Une modélisation fine de la génératrice sur la méthode des circuits internes équivalents a été réalisée et un algorithme de calcul des paramètres à partir des données géométriques de la machine a été développé permettant d'automatiser le calcul pour n'importe quels stators et schémas de bobinage. Associé au convertisseur, ce modèle a été simulé avec succès et une commande vectorielle a également été introduite à ce schéma. Cette stratégie de contrôle permet d'adapter les matrices de transformation ainsi que les paramètres des régulateurs PI en fonction du défaut et confère une tolérance aux défauts électriques. Cette adaptation permet de réduire significativement les oscillations de puissance lors de la perte d'une ou plusieurs phases. Pour valider les théories développées et déjà simulées, des essais ont été réalisés avec succès sur un banc d'essai de 24kW, image d'une éolienne connectée au réseau / For around 15 years, wind turbines have found a wide popularity and increase in terms of number and power per unit but they have still to deal with mechanical and electrical faults. Then, the aim of this thesis is to design, model and control a wind turbine generator that is able to cope with these problems. For this, a structure based on a squirrel cage induction machine with 6 phases and 24 poles has been studied. Indeed, by increasing the number of poles, one can simplify or eliminate the gearbox that induces many faults while a multiphase structure allows electrical energy production when several stator phases or inverter legs are lost. For this, a precise model of the generator has been developed using the equivalent intern circuits and a parameters computing strategy that allows the determination of the parameters whatever the geometrical and electrical structure of the stator has been introduced. Associated to the power converter, this model has been simulated successfully and a field oriented control has also been inserted in the whole simulation scheme. This control strategy allows tuning of the transformation matrices and also PI regulators parameters as function of the fault and therefore is robust against electrical parameters changes. Indeed, the on-line adaptation lets to reduce significantly the power ripples that appear when one or more phases are lost. To validate the proposed method that have been previously simulated, the same test have been carry out successfully on a 24 kW prototype that is a picture, at scale 1/100, of a real advanced wind turbine connected to the grid

This thesis presents a performance evaluation of a micro horizontal axis wind turbine, investigates the use of particle image velocimetry (PIV) to capture the flow field around a rotating blade and to track water droplets in the flow. The testing was done in a low speed wind tunnel in a highly blocked configuration. The turbine was instrumented to measure rotational speed of the rotor, axial thrust and power output. Wind speed of the wake was measured with a Kiel probe. Performance characteristics were calculated and compared with the manufacturer’s published data and to power predictions by axial momentum theories. The turbine was shown to perform well and the manufacturer’s published data are accurate. Axial momentum theory over-predicts power by approximately 50%. It is shown that good PIV results can be obtained using a fog machine to seed the flow. Improved illumination and optics will be required to measure 3D flow close to the blade. Water droplets can be tracked but a shadowgraphy arrangement should be used to better visualize the droplets. The droplets also affect the rotational speed of the rotor such that capturing the blade in a consistent point in the field of view is problematic.

<p>Power generation using wind turbines is a highly researched control field.</p><p>Many control designs have been proposed based on continuous-time models</p><p>like PI-control, or state observers with state feedback but without special</p><p>regard to robustness to model uncertainties. The aim of this thesis was to</p><p>design a robust digital controller for a wind turbine.</p><p>The design was based on a discrete-time model in the polynomial framework</p><p>that was derived from a continuous-time state-space model based on</p><p>data from a real plant. A digital controller was then designed by interactive</p><p>pole placement to satisfy bounds on sensitivity functions.</p><p>As a result the controller eliminates steady state errors after a step</p><p>response, gives sufficient damping by using dynamical feedback, tolerates</p><p>changes in the dynamics to account for non linear effects, and avoids feedback</p><p>of high frequency un modeled dynamics.</p>

As international growth in wind energy steadily increases and the world gradually moves away from fossil fuels, advanced computational tools are required to produce accurate and fast predictions in wind turbine performance, and to allow efficient design cycles using advanced materials and manufacturing methods. Currently, aerostructural analysis often employs the relatively fast but inaccurate Blade Element Momentum (BEM) theory, while accurate but slower Computational Fluid Dynamics (CFD) methods are generally used for aerodynamic analysis alone.To bridge the gap between speed and accuracy, a 3D panel code, TriPan, was coupled with an advanced structural Finite Element Method (FEM) code, TACS, to perform aerostructural analysis for wind turbine blades. In addition, the framework allows the replacement of the panel solver by higher fidelity solvers to increase the accuracy of the overall aerostructural solution.

As international growth in wind energy steadily increases and the world gradually moves away from fossil fuels, advanced computational tools are required to produce accurate and fast predictions in wind turbine performance, and to allow efficient design cycles using advanced materials and manufacturing methods. Currently, aerostructural analysis often employs the relatively fast but inaccurate Blade Element Momentum (BEM) theory, while accurate but slower Computational Fluid Dynamics (CFD) methods are generally used for aerodynamic analysis alone.To bridge the gap between speed and accuracy, a 3D panel code, TriPan, was coupled with an advanced structural Finite Element Method (FEM) code, TACS, to perform aerostructural analysis for wind turbine blades. In addition, the framework allows the replacement of the panel solver by higher fidelity solvers to increase the accuracy of the overall aerostructural solution.

Wind energy is the most abundant resource in the renewable energy portfolio. Increasing the wind capture capability improves the economic viability of this technology, and makes it more competitive with traditional fossil-fuel based supplies. Therefore, it is necessary to explore control strategies that maximize aerodynamic efficiency, thus, the wind energy capture. Several control algorithms are developed and compared during this research. A traditional feedback control is adapted as the benchmark approach, where the turbine torque and the blade pitch angle are used to control the wind turbine operation during partial and full load operations, correspondingly. Augmented feedback control algorithms are then developed to improve the wind energy harvesting. Optimal control methodologies are extensively explored to achieve maximal wind energy capture. Numerical optimization techniques, such as direct shooting optimization are employed. The direct shooting method convert the optimal control problem into a parameter optimization problem and use nonlinear programming algorithm to find the optimal solution. The dynamic programming, a global optimization approach over a time horizon, is also investigated. The dynamic programming finds the control inputs for the blade pitch angle and speed ratio to maximize the power coefficient, based on historical wind data. A dynamic wind turbine model has been developed to facilitate this process by characterizing the performance of the various possible input scenarios. Simulation results of each algorithm on real wind site data are presented to compare the wind energy capture under the proposed control algorithms with the traditional feedback control design. The result of the tradeoff analysis between the computation expense and the energy capture is also reported. / text

The growth in the number of wind farms has raised significant concerns in the radar community due to their potential interference on radar systems. The motion of the turbine blades creates unwanted Doppler clutter that can interfere in the tracking of moving targets. Large turbine structures can also produce electromagnetic shadows that may make observing objects behind a wind farm difficult. Detailed characterization of the clutter is the first step towards effective mitigation techniques. The goal of this dissertation research is to gain a better understanding of the dynamic radar signatures resulting from scattering by wind turbines. First, the scattering characteristics of turbines in the presence of ground surface are studied. Image theory in conjunction with a shooting-and-bouncing ray code, Ahilo, is used to carry out the dynamic signature simulation. The observed features in the simulation are corroborated with laboratory model measurements. Second, the effects of higher order motions of a turbine undergoing rotation on the radar signatures are investigated and characterized. Mathematical models for the motions are proposed and used to simulate the joint time-frequency and inverse synthetic aperture radar characteristics of the turbine undergoing these motions. The motions are studied for an isolated turbine as well as for a turbine rotating above a ground. Selected motions are corroborated by laboratory model measurements. Next, a method to remove the dynamic clutter produced by wind turbines is presented. A physics-based basis is constructed to model the radar backscattering from a wind turbine. This basis is used in conjunction with the matching pursuit algorithm to iteratively remove the Doppler clutter due to wind turbines. The algorithm is tested using radar return generated using Ahilo. Finally, radar features of wind turbines are simulated and studied in the HF (high frequency) band. The features are presented in the range-Doppler plane for single as well as arrays of turbines. Doppler aliasing due to the limited pulse repetition frequency of HF radars is examined. Shadowing characteristics of arrays of turbines are simulated and analyzed. Electromagnetic modeling details including effects of thin-wire modeling, non-conducting turbine components, and the presence of a conducting ground surface are discussed. / text

Power generation using wind turbines is a highly researched control field. Many control designs have been proposed based on continuous-time models like PI-control, or state observers with state feedback but without special regard to robustness to model uncertainties. The aim of this thesis was to design a robust digital controller for a wind turbine. The design was based on a discrete-time model in the polynomial framework that was derived from a continuous-time state-space model based on data from a real plant. A digital controller was then designed by interactive pole placement to satisfy bounds on sensitivity functions. As a result the controller eliminates steady state errors after a step response, gives sufficient damping by using dynamical feedback, tolerates changes in the dynamics to account for non linear effects, and avoids feedback of high frequency un modeled dynamics.

No description available.

The focus of the research program is to analysis and design of a new generation glass-fibre reinforced polymer (GFRP) wind turbine tower for full scale prototype testing. The study includes the finite element analyses of different tower section configurations, the parametrical study of different variables, the selection of appropriate configuration and dimensions, and the finalization of the section. The design section arrived from this study has the bottom outer diameter of 1350 mm, the top outer diameter of 800 mm, the constant inner diameter of 600 mm and uniform wall thickness of 11.25 mm. The tower is also analysed and compared with a steel tapered tower. The analysis results indicate that the tower is considered as a soft-soft tower and that, in general, the lateral deflection limitation is a governing factor in the design of GFRP wind turbine tower. The proposed section met all the design requirements and the fabrication drawings are provided for the further study of full scale test.