Au cours des dernières décennies, la possibilité d’exploiter les capacités de vol exceptionnelles des insectes a été à l’origine de nombreuses recherches sur l’élaboration de nano-véhicules aériens (NAVs) à ailes battantes. Cependant, lors de la conception de tels prototypes, les chercheurs doivent analyser une vaste gamme de solutions liées à la grande diversité des insectes volants pour identifier les fonctionnalités et les paramètres adaptés à leurs besoins. Afin d’alléger cette tâche, le but de ce travail est de développer un outil permettant à la fois d’examiner le comportement cinématique et énergétique d’un nano-véhicule aérien à ailes flexibles résonantes, et donc d'évaluer son efficacité. Cet objectif reste néanmoins extrêmement difficile à atteindre car il concerne des objets de très petites tailles. Aussi, nous avons choisi tout d’abord de travailler sur un micro-véhicule aérien (MAV) à ailes battantes. Il s’agit avant tout de valider l’outil de modélisation à travers une comparaison systématique des simulations avec des résultats expérimentaux effectués lors de l’actionnement des ailes, puis au cours du décollage et du vol stationnaire du prototype. Une partie des connaissances et expériences acquises pourra ensuite être utilisée afin de mieux comprendre le fonctionnement et identifier la distribution d'énergie au sein du NAV. Bien que les deux véhicules s’inspirent directement de la cinématique des ailes d'insectes, les mécanismes d'actionnement des ailes artificielles des deux prototypes ne sont pas les mêmes en raison de la différence de taille. Comme le NAV est plus petit, ces ailes ont un mouvement de battement à une fréquence plus élevée que celles du MAV, à l’instar de ce qui existe dans la nature. En conséquence, lorsque l’on passe du MAV au NAV, le mécanisme d’actionnement des ailes doit être adapté et cette différence nécessite d’une part, de revoir la conception, l'approche de modélisation et le processus d'optimisation, et d’autre part, de modifier le procédé de fabrication. Une fois ces améliorations apportées, nous avons obtenu des résultats de simulations en accord avec les tests expérimentaux. Le principal résultat de ce travail concerne l’obtention pour les deux prototypes, le MAV et le NAV, d’une cinématique appropriée des ailes, qui conduit à une force de portance équivalente au poids. Nous avons d’ailleurs démontré que le MAV était capable de décoller et d’avoir un vol stationnaire stable selon l’axe vertical. En tirant parti des modèles basés sur le langage Bond Graph, il est également possible d'évaluer les performances énergétiques de ces prototypes en fonction de la dynamique de l'aile. En conclusion, cette étude contribue à la définition des paramètres essentiels à prendre en compte lors de la conception et l'optimisation énergétique de micro et nano-véhicules à ailes battantes. / In recent decades, the prospect of exploiting the exceptional flying capacities of insects has prompted much research on the elaboration of flapping-wing nano air vehicles (FWNAV). However, when designing such a prototype, designers have to wade through a vast array of design solutions that reflects the wide variety of flying insects to identify the correct combination of parameters to meet their requirements. To alleviate this burden, the purpose of this work is to develop a suitable tool to analyze the kinematic and power behavior of a resonant flexible-wing nano air vehicle. The key issue is evaluating its efficiency. However, this ultimate objective is extremely challenging as it is applied to the smallest flexible FWNAV. However, in this work, we worked first with a flapping-wing micro air vehicle (FWMAV) in order to have a tool for the simulation and experimentation of wing actuation, take-off and hovering. Some of the knowledge and experience acquired will then be transferred to better understand how our FWNAV works and identify the energy, power distribution. Although both of the vehicles employ the insect wing kinematics, their wings actuation mechanisms are not the same due to their sizes difference. Since the FWNAV is smaller, their wings flap at a higher frequency than the FWMAV as inspired by nature. As a consequence, from MAV to NAV, the wing actuation mechanism must be changed. Throughout this work, it can be seen clearly that this difference affects the whole vehicles development including the design, the manufacturing method, the modeling approach and the optimizing process. It has been demonstrated that the simulations are in good correlation with the experimental tests. The main result of this work is the proper wing kinematics of both FWMAV and FWNAV which leads to a lift to the weight ratio bigger and equal to one respectively. The FWMAV is even success to take-off and vertically stable hover. Moreover, taking advantage of the Bond Graph-based models, the evolution power according to the wing dynamic and the efficiency of the subsystem can be evaluated. In conclusion, this study shows the key parameters for designing and optimizing efficiency and the lift generated for two flapping wing vehicles in different size regimes.

This thesis contributes to the state of the art in integrated design of insect-scale piezoelectric actuated flapping wing vehicles through the development of novel theoretical models for flapping wing aerodynamics and piezoelectric actuator dynamics, and integration of these models into a closed form design process. A comprehensive literature review of available engineered designs of miniature rotary and flapping wing vehicles is provided. A novel taxonomy based on wing and actuator kinematics is proposed as an effective means of classifying the large variation of vehicle configurations currently under development. The most successful insect-scale vehicles developed to date have used piezoelectric actuation, system resonance for motion amplification, and passive wing pitching. A novel analytical treatment is proposed to quantify induced power losses in normal hover that accounts for the effects of non uniform downwash, wake periodicity and effective flapping disc area. Two different quasi-steady aerodynamic modelling approaches are undertaken, one based on blade element analysis and one based on lifting line theory. Both approaches are explicitly linked to the underlying flow physics and, unlike a number of competing approaches, do not require empirical data. Models have been successfully validated against experimental and numerical data from the literature. These models have allowed improved insight into the role of the wing leading-edge vortex in lift augmentation and quantification of the comparative contributions of induced and profile drag for insect-like wings in hover. Theoretical aerodynamic analysis has been used to identify a theoretical solution for the optimum planform for a flapping wing in terms of chord and twist as a function of span. It is shown that an untwisted elliptical planform minimises profile power, whereas a more highly tapered design such as that found on a hummingbird minimises induced power. Aero-optimum wing kinematics for hovering are also assessed. It is shown that for efficient flight the flapping velocity should be constant whereas for maximum effectiveness the flapping velocity should be sinusoidal. For both cases, the wing pitching at stroke reversal should be as rapid as possible. A dynamic electromechanical model of piezoelectric bending actuators has been developed and validated against data obtained from experiments undertaken as part of this thesis. An expression for the electromechanical coupling factor (EMCF) is extracted from the analytical model and is used to understand the influence of actuator design variables on actuator performance. It is found that the variation in EMCF with design variables is similar for both static and dynamic operation, however for light damping the dynamic EMCF will typically be an order of magnitude greater than for static operation. Theoretical contributions to aerodynamic and electromechanical modelling are integrated into a low order design method for propulsion system sizing. The method is unique in that aside from mass fraction estimation, the underlying models are fully physics based. The transparency of the design method provides the designer with clear insight into effects of changing core design variables such as the maximum flapping amplitude, wing mass, transmission ratio, piezoelectric characteristics on the overall design solution. Whilst the wing mass is only around 10% of the actuator mass, the effective wing mass is 16 times the effective actuator mass for a typical transmission ratio of 10 and hence the wing mass dominates the inertial contribution to the system dynamics. For optimum aerodynamic effectiveness and efficiency it is important to achieve high flapping amplitudes, however this is typically limited by the maximum allowable field strength of the piezoelectric material used in the actuator.