The detailed behaviour of the unsteady boundary layer and stall events occurred on a sinusoidally oscillating NACA 0012 airfoil at Re = 1.35 x 105 was investigated experimentally by using closely-spaced multiple hot-film sensor arrays. The hot-film measurements were supplemented by surface pressure measurements, hot-wire wake velocity surveys and smoke-flow visualizations. Three typical oscillation parameters: attached flow, light stall, and deep stall were tested. Special attention was focused on the non-intrusive identification of the spatial-temporal progression of the locations of the boundary-layer transition and separation and reattachment points for a range of oscillation frequencies and amplitudes both prior to, during, and post the stall. The results show that for an unsteady airfoil, the reduced frequency of the oscillation was found to be highly significant and only small values of reduced frequency were required to delay the onset of the various boundary-layer events, and to produce significant variations and hysteresis in the peak values of lift and drag forces and the pitching moments, which are fundamentally different from their static counterparts. Lift stall was observed to occur when the leading-edge vortex reached 90% of the chord, while moment stall occurred at the end of the upward spread of the trailing-edge flow reversal. The convection speed of the leading-edge and the secondary vortices were also reported. Dynamic stall was found to be caused by an abrupt turbulent separation near the leading-edge region and not with the bursting of the laminar separation bubble, as is commonly observed for a static NACA 0012 airfoil; the result being the initiation, growth and convection of an energetic leading-edge vortex. Moreover, the detection of the sudden turbulent breakdown could serve as an indicator for dynamic stall detection and control. The results reported on here provide a deepened insight into the detailed nature of

Increasing the combustor exit temperature in gas turbines is an effective means to increase engine power. While occasional metallurgical advances allow gradual temperature increases, improving the internal/external cooling of the blades is the only way to permit significant temperature gains. In this work, a methodology for aerodynamic and conjugate heat transfer computational analysis of cooled turbine blades is developed. Flow solutions are obtained using an implicit, three-dimensional, finite-element Reynolds-Averaged Navier-Stokes flow solver. Efficient non-reflecting boundary conditions are derived and implemented to reduce the size of the solution domain and accelerate convergence. These are shown to be essential for the accurate capturing of shock waves and wakes. The methodology is demonstrated on the convection-cooled NASA-C3X turbine vane, by coupling heat conduction in the solid vane with heat transfer from the internal cooling flow and the external hot-gas flow. Both aerodynamic and heat transfer results are compared against experimental data.

In aviation, synthetic vision systems render artificial views of the world (using a database of the world and pose information) to support navigation and situational awareness in low visibility conditions. The database needs to be periodically updated to ensure its consistency with reality, since it reflects at best a nominal state of the environment. This thesis presents an approach for automatically updating the geometry of synthetic vision system databases and 3D models in general. The approach is novel in that it profits from all of the available prior information: intrinsic/extrinsic camera parameters and geometry of the world. Geometric inconsistencies (or anomalies) between the model and reality are quickly localized; this localization serves to significantly reduce the complexity of the updating problem. Given a geometric model of the world, a sample image and known camera motion, a predicted image can be generated based on a differential approach. Model locations where predictions do not match observations are assumed to be incorrect. The updating is then cast as an optimization problem where differences between observations and predictions are minimized. To cope with system uncertainties, a mechanism that automatically infers their impact on prediction validity is derived. This method not only renders the anomaly detection process robust but also prevents the overfitting of the data. The updating framework is examined at first using synthetic data and further tested in both a laboratory environment and using a helicopter in flight. Experimental results show that the algorithm is effective and robust across different operating conditions.

Statistical inhomogeneity of material properties, variations in nominal geometry, manufacturing tolerances, operational wear lead to uncertainties in the parameters associated with FE models of turbine engine rotors and consequently to uncertainties in their vibration response. Reliable assessment of the rotor system behavior cannot be made unless the effects of such uncertainties are understood and quantified. / In practical situations the parametric probabilistic approach is the first choice to employ in that context yielding efficient algorithms with feasible implementations. A set of measured or estimated experimentally random parameters is repeatedly propagated through rotor models in Monte-Carlo simulations, which would pose a formidable computational task if the full order high-fidelity finite element (FE) models were utilized. The objective of this dissertation is to decrease the expense of analyzing systems modified in the parametric space by developing accurate model reduction computational techniques suitable for repeated analysis, in particular addressing the problem of large variations in nominal geometry. / The first part of the dissertation is concerned with the structural blade mistuning problem. The existing projection based model order reduction techniques capable to numerically characterize variations in nominal geometry of periodic structures are examined, a method generating very compact reduced order models (ROM) based on correction, as opposed to expansion, of the modal subspace is selected and its limitations are analyzed. A new algorithm drawing on optimal preconditioned iterative methods for generalized eigenvalue problem is introduced to address its deficiencies. Both techniques are combined in a stochastic simulation framework to analyze the effect of random mistuning on geometrically modified bladed disks, where random parameter variation in blade properties is introduced in modal space at component level. A family of benchmark problems on an industrial scale bladed disk model are utilized in a comparative study assessing the amount of computational effort and storage, scalability and accuracy as well as providing insight on underlying physical phenomena. / In the second part of dissertation a new computational technique is proposed focusing on prediction of the effects of uncertainty in rotor assembly inter-stage geometry on global vibration response. The algorithm stands apart from the traditional modal projection based framework employing harmonic truncation only. It is shown that decent performance can be achieved due to reliance on sparse matrix linear algebra and sampling of small parametric space. Particular emphasis is given to the computational efficiency of ROM update. Accuracy and performance of the technique is illustrated with representative simulation examples over a practical range of geometrical parameter variations and operational conditions. / La prise en compte des effets des incertitudes est un élément fondamental pour une conception fiable des machines tournantes. Les inhomogénéités matérielles, les variations de géométrie, les tolérance de fabrication ou encore les phénomènes d'usure en service comptent parmi les nombreuses sources de variabilité des paramètres de modélisation qui conduisent, au niveau de la réponse dynamique des structures, à des incertitudes qu'il est nécessaire de quantifier. / En pratique, l'approche probabiliste paramétrique est souvent privilégiée pour sa facilite des mise en oeuvre ainsi que pour l'efficacité des algorithmes utilisables. Dans le cadre de simulations de Monte-Carlo, qui consistent à évaluer de façon répétée un modelé en fonction d'un grand nombre de réalisations de paramètres aléatoires (détermines expérimentalement ou estimes), les coûts de calcul peuvent s'avérer prohibitifs si des modelés éléments-finis de haute fidélité sont utilises. L'objectif de cette thèse est de diminuer l'effort nécessaire à l'évaluation de la réponse de systèmes incertains en développant des techniques de réduction de modelés adaptées à des analyses répétées et permettant, en particulier, la prise en compte de grandes variations géométriques. / La première partie de cette dissertation concerne le problème du désaccordage structurel des roues aubagées. Les techniques existantes de réduction de modelés basées des projections qui permettent de caractériser les variations de géométrie de structures périodiques sont dans un premier temps revues. Une méthode amenant un modelé réduit très compact, basée sur la correction et non l'expansion du sousespace modal est ensuite sélectionnée et ses limitations sont analysées. Un nouvel algorithme, inspiré des méthodes itératives de pré-conditionnement optimal pour les problèmes aux valeurs propres généralises, est ensuite introduit pour palier ces insuffisances. Les deux techniques sont combinées dans le cadre de simulations sophistiqués pour analyser les effets désaccord aléatoire et de modifications géométriques pour les roues aubagees. Les variations aléatoires sont introduites en tant que paramètres des aubes dans l'espace modal. Une étude comparative est ensuite présentée en s'appuyant sur un modelé de taille industrielle. Les coûts de calcul, les possibilités d'extension ou encore la précision des méthodes sont examines et des aspects phénoménologie sont discutes. / Dans la second partie de cette dissertation, une nouvelle approche numérique est proposée pour la quantification des effets d'incertitudes géométriques dans les assemblages inter-étages des rotors. L'algorithme se distingue des approches traditionnelles basées sur des projections en cela que seule une troncature harmonique est considérée. Des performances satisfaisantes peuvent être obtenues grâce à des opérations d'algèbre linéaire sur des matrices creuses ainsi qu'à l'échantillonnage d'un espace paramétrique minimal. En particulier, l'efficacité numérique de la mise a jour du modelé réduit sera mise en évidence. Enfin, la précision et le performance de cette méthode seront illustrées au moyen d'exemples représentatifs tant en termes de variations de paramètres géométriques que de conditions de fonctionnement.

The research described in this thesis relates to a highly-maneuverable, almost-lighter-than-air vehicle (ALTAV). The vehicle is controlled by four vectored thrusters and is marginally stable due to its finless design. In order to better understand the behavior of this vehicle, a non-linear mathematical model was developed to represent the behavior of the airship under the effect of thruster forces and wind. The underlying equations of motion in the dynamics model take into account aerostatic and aerodynamic forces on the airship. Physical parameters used in the equations of motion were estimated through experiments and a CAD model. The viscous drag acting on the hull was computed for a 360 degrees range of angle of attack by adapting an existing semi-empirical method for slender bodies. A model of the vectored thrusters developed from experimental data was shown to provide good agreement with actual thrust measurements under static conditions. Reference flight data for model validation was obtained by measuring the airship's response to environmental disturbances and thruster inputs. In general, the simulation was found to provide a reasonable estimate of the airship's roll, pitch and vertical trajectory. However, the airship motion was strongly affected by wind. This suggests that better results would be obtained if the simulation could better reproduce the wind conditions existing during the flight tests. / La recherche décrite dans cette thèse fait réference à un vehicule flottant très manœuvrable. Le véhicule est contrôlé par quatre propulseurs et a peu de stabilité en raison de sa conception sans ailettes. Afin de mieux comprendre le comportement de ce véhicule, un modèle mathématique non-linéaire a été développé pour representer le comportement de l'aéronef sous les effets de ses propulseurs et du vent. Les équations de déplacements/mouvements dans le modèle dynamique prennent en considération les interactions aérostatiques et aérodynamiques qui se produisent entre l'aéronef et l'air environnant. Les paramètres physiques utilisés dans les équations de mouvement ont été estimés à l'aide déssais réels et d'un modèle CAO(CAD). La resistance aérodynamique agissant sur le châssis a été calculée pour des angles d'attaque de 0 deg à 360 deg en adaptant une méthode semi-empirique applicable aux corps minces. Un modèle des propulseurs développé à partir de données expérimentales a démontré un bon accord avec les données de propulsion réelles dans des conditions statiques. Les données pour la validation du modeèle ont été obtenues en mesurant la reaction de l'aéronef sous léffet de perturbations environnementales et de l'action des propulseurs. En général, la simulation a fourni une estimation raisonnable du roulis, du tangage, et de la trajectoire verticale de l'aéronef. Toutefois, le mouvement de l'aéronef est susceptible d'être fortement affecté par le vent. Ceci suggère que de meilleurs résultats seraient obtenus si la simulation pouvait mieux reproduire les conditions de vent existantes pendant les essais en vol.

In-flight icing is a hazard that continues to afflict the aviation industry, despite all the research and efforts to mitigate the risks. The recurrence of these types of accidents has given renewed impetus to the development of advanced analytical predictive tools to study both the accretion of ice on aircraft components in flight, and the aerodynamic consequences of such ice accumulations. In this work, an in-depth analysis of the occurrence of in-flight icing accidents and incidents was conducted to identify high-risk flight conditions. To investigate these conditions more thoroughly, a computational fluid dynamics model of a representative airfoil was developed to recreate experiments from the icing wind tunnel that occurred in controlled flight conditions. The ice accumulations and resulting aerodynamic performance degradations of the airfoil were computed for a range or pitch angles and flight speeds. These simulations revealed substantial performance losses such as reduced maximum lift, and decreased stall angle. From these results, an icing hazard analysis tool was developed, using risk management principles, to evaluate the dangers of in-flight icing for a specific aircraft based on the atmospheric conditions it is expected to encounter, as well as the effectiveness of aircraft certification procedures. This method is then demonstrated through the simulation of in-flight icing scenarios based on real flight data from accidents and incidents. The risk management methodology is applied to the results of the simulations and the predicted performance degradation is compared to recorded aircraft performance characteristics at the time of the occurrence. The aircraft performance predictions and resulting risk assessment are found to correspond strongly to the pilot's comments as well as to the severity of the incident. / Le givrage en cours de vol constitue un danger qui continue d'affliger l'industrie aéronautique. Ceci, malgré toute la recherche qui vise à atténuer les risques. La récidive de ce type d'accident donna un nouvel élan au développement d'outils analytiques de pointe dans ce domaine. Ces outils permettent d'étudier de manière préventive, l'accrétion de glace sur les composants de l'avion, ainsi que les conséquences aérodynamiques de telles accumulations. Le présent travail comprend une analyse approfondie de l'occurrence d'incidents liés au givrage en cours de vol dans le but d'identifier les conditions météorologiques problématiques. Pour en apprendre d'avantage, un modèle représentatif d'une aile d'avion à haut risque fut développé. Les conditions d'opérations de cette aile furent simulées par un logiciel d'analyse de dynamique des fluides afin de recréer les résultats de simulations menées en soufflerie givrante. La quantité d'accumulation de glace ainsi que la dégradation des performances aérodynamiques sont calculées pour une gamme de vitesse et d'angles de tangage. Ces simulations démontrent une perte importante de la portance maximale de l'aile ainsi qu'une réduction de l'angle de décrochage. À partir de ces résultats, une méthodologie d'analyse de risques fut développée, selon les principes de gestion de sureté. Le but d'une telle analyse est d'évaluer le niveau de danger associé au givrage en vol pour un appareil précis selon des conditions de vol déterminées avant le décollage. De plus, cette méthodologie servira à évaluer l'efficacité du système de dégivrage, ainsi que le protocole de certification d'un appareil qui opère dans des conditions propice au givrage en vol. Le fonctionnement du système de gestion de sécurité est ensuite démontré à l'aide de simulations d'accidents liés au givrage en cours de vol selon les informations obtenues dans le rapport d'incident. La méthodologie fut ensuite appliquée aux résultats des simulations; les pertes de performance, calculées par le logiciel d'analyse de dynamique des fluides, furent comparées à la performance et au comportement de l'aéronef tel qu'enregistré au moment de l'accident. Les résultats démontrent que le système de gestion de risque développé dans cet ouvrage peut prédire de façon précise, avec l'information disponible uniquement avant le décollage, la sévérité du givrage en vol pour un avion spécifique.

Tethers have been used in space for quite some time. However, the concept of a multi-tethered system has received less attention. This thesis investigates the dynamics of certain configurations of multi-tethered satellite formations in Low Earth Orbits. / Certain simplifications have been made prior to the investigation of the problem; these include considering the satellites to be point-masses and the tethers to be massless and straight. Configurations termed hub-and-spoke and closed-hub-and-spoke are analyzed. The local motion is studied under two distinct specifications for the orbital motion of the system: first, prescribed motion of the system centre of mass and second, prescribed motion of the central body of the formation. Three-dimensional motion is studied for formations lying in the orbital plane and in the Earth-facing plane normal to the orbit, while given an initial spin rate in their nominal plane. / Later, two satellite formations deployed along the local vertical are examined: a closed-hub-and-spoke configuration with its spin axis along the long axis of symmetry, which is found to be unstable, and a double-pyramid configuration, which is found to be stable.

1AbstractCurved shock theory (CST) has been extended to apply to axisymmetric shocks in non-uniform flow. A general formula has been derived for the vorticity jump across a doubly curved shock in non-uniform flow. Influence coefficient forms of equations for the gradients and vorticity show the effect of changing pre-shock conditions. CST has been applied to a series of simple shock flows and to the orientation of the sonic surface at the rear face of a doubly curved shock. This orientation is significant in determining the occurrence of embedded shocks in the post-shock flow. Application of CST to curved, concave, normal shocks allowed the derivation of an explicit relationship between the shock's curvature and the length of down-shock subsonic flow. Investigations of conical flows by analysis, CFD and experiment all failed to demonstrate the existence of regular reflection of shocks at the centre line of axisymmetric flows. An analytically predicted conical shock, on the calculated streamline, does not extend all the way to the centre line but terminates in Mach reflection. It appears that the existence of an analytical Taylor-Mccoll (T-M) solution is not in itself a guarantee of the physical existence of a conical flow in all cases. The T-M equations predict the existence of an axisymmetric centered compression fan, analogous to the Prandtl-Meyer fan in planar flow. A free-standing conical shock is located downstream of the compression fan. Both features have been shown to exist by CFD as well as experiment. Busemann flow is the only flow where these wave structures can exist; it is possible to reflect an incident, centered compression as a conical shock. Discovery of an inflection point on the Busemann streamline has an important implication to spontaneous starting of Busemann intakes. Three types of flow can exist behind a doubly curved concave shock; characterized by the orientation of the sonic surface which, in turn, is determined by the pre-shock Mach number and the shock curvatures ratio. Shapes of special axial shock surfaces, with straight post shock streamlines (Crocco shocks), or vanishing streamwise pressure gradient (Thomas shocks) and shocks with specific sound reflectivity (zero, if desired), have been calculated and illustrated. Boundary layer generated noise abatement is a possibility. Local flow choking, near the leading edge, leads to shock detachment from a curved wedge with such detachment depending on freestream Mach number, the wedge2angle, the wedge curvature and the wedge length. These are new criteria for shock detachment with analogies extending to the transition from regular to Mach reflection of shock waves. / 1RésuméLa théorie des ondes de chocs courbées (TOCC; Curved Shock Theory) a été généralisée aux chocs axisymétriques dans un écoulement non uniforme. Une formule générale a été dérivée pour les sauts de vorticité à travers un choc à double courbe dans un écoulement non uniforme. La forme coefficient d'influence des équations des gradients et de la vorticité démontrent l'effet de la variation des conditions en amont. La TOCC a été appliquée à plusieurs écoulements simples avec chocs incluant l'orientation de la surface sonique à la face arrière d'un choc à double courbe. Cette orientation est importante pour déterminer l'existence d'ondes de choc intégrées à l'écoulement aval. L'application de la TOCC aux ondes de choc courbées, concaves et normales permet de dériver une relation explicite entre la courbe du choc et la longueur de l'écoulement subsonique derrière l'onde. L'étude analytique, numérique et expérimentale des écoulements coniques n'a pas permis de démontrer l'existence de réflexions régulières des chocs à l'axe de symétrie des écoulements. Un choc conique prédit analytiquement sur la ligne d'écoulement n'atteint pas l'axe central, mais se termine en réflexion Mach. Il semble que l'existence d'une solution Taylor-Mccoll (T-M) ne garantit pas l'existence physique d'un écoulement conique. Les équations T-M prédisent l'existence d'un train d'ondes de compression axisymétrique, analogue au train d'ondes de Prandtl-Meyer dans un écoulement planaire. Un choc conique détaché est situé en aval du train de compression. L'existence des deux caractéristiques a été démontrée par CFD ainsi qu'expérimentalement. L'écoulement Busemann est le seul écoulement où ces structures d'ondes peuvent exister : une compression centrée peut être reflétée en onde de choc conique. La découverte d'un point d'inflexion dans la ligne d'écoulement de Busemann a une implication importante au démarrage spontané de diffuseurs Busemann. Trois types d'écoulements peuvent exister à l'arrière d'un choc concave à double courbure : ils sont caractérisés par l'orientation de la surface sonique qui, à son tour, est déterminée par le nombre de Mach pré-choc et le ratio de courbures du choc. Des formes de surfaces d'ondes de choc axiales particulières, avec écoulement droit en aval (chocs Crocco), ou avec un gradient de pression tendant vers zéro dans l'axe d'écoulement (chocs Thomas) ainsi que des chocs avec une réflectivité acoustique spécifique (incluant nulle) ont été calculées et illustrées. Une réduction du bruit de couche limite est aussi possible.2L'étranglement local au bord d'attaque d'une pointe courbée mène au détachement de l'onde de choc, lequel dépend du nombre de Mach de l'écoulement libre, de l'angle, de la courbure et de la longueur de la pointe. Ce sont de nouveaux critères pour le détachement du choc avec des analogies pouvant s'étendre aux transitions des réflexions régulières aux réflexions Mach.

The main objective of this thesis is to investigate the effects of turbulent flow on the random dynamics of structurally nonlinear airfoil. A secondary objective is to articulate a more comprehensive picture of the contribution of the longitudinal component of turbulence, as experienced by the airfoil, be it linear or nonlinear. In this regard, a systematic and detailed numerical analysis of the airfoil experiencing random flutter/Hopf bifurcation is presented. Some aspects of the divergence/pitchfork problem are also discussed. / The airfoil is modelled as a flexibly mounted rigid flat plate with degrees-of-freedom in pitch and heave. The principal nonlinearity considered is a hardening cubic torsional spring. The aerodynamics is incompressible and linear. Unsteady aerodynamic effects due to arbitrary motion and turbulence are modelled. Both longitudinal and vertical components of the Gaussian turbulence are considered. Longitudinal turbulence acts as a parametric excitation, whereas the latter represents an external forcing. / A Monte Carlo simulation is performed to solve numerically the system of random differential equations. The time history solutions are then studied in terms of their mean-square, probability density function and power spectral density. The largest Lyapunov exponent is also calculated. / The bifurcation, stability and response characteristics of the airfoil are examined. For the linear airfoil, it is found that the coalescence flutter speed is always advanced by the longitudinal component of turbulence, and generally dominated by the very low frequency range of the excitation. Divergence can be either advanced or postponed, but the magnitude of the shift is not significant compared with flutter. Furthermore, it is shown that in general longitudinal turbulence decreases the overall stability of the airfoil, be it linear or nonlinear. / For the nonlinear airfoil, it is the vertical component of turbulence that determines the essential features of the stochastic bifurcation and the qualitative characteristics of the response. The interplay between turbulence and nonlinear stiffness has a significant impact on the probability structure of the aeroelastic response. Uni-, bi- and double bi-modal distributions are observed, and found to occur at different airspeeds depending on which state variable is considered. Furthermore, the spectral content displays noise-controlled, and noise-induced, time scales.

The near-field flow structure of the tip vortex generated by a NACA 0015 wing oscillating though the attached-flow, light-stall and deep-stall regimes was investigated at Rec = 1.86 x 105. Particular emphasis was placed on the effects of oscillation frequency and mean incidence upon the spatial and temporal evolution of the unsteady vortex structure. Phase-locked, ensemble-averaged cross-flow and axial velocity fields, vorticity distributions, and turbulence structures over a full cycle of oscillation were compared to static wing-tip vortex results, and the dynamic effects upon the vortex strength, size, trajectory and associated induced drag were examined. Through the attached-flow and light-stall oscillations, most vortex properties were qualitatively similar to the static cases, though a small degree of hysteresis between the pitch-up and pitch-down phases of motion was observed. The radial distributions of circulation within the inner region of the vortex were self-similar, and showed only small variations from the static case. When the wing was oscillated through the deep-stall regime, a dramatic decrease in tip vortex strength and concentration was observed at the end of the upstroke, as a result of the growth of the leading-edge vortex and subsequent catastrophic flow separation. The use of passive spoilers and active flaps to control the strength and trajectory of the tip vortex was also investigated.