No description available.

Dans les centrales nucléaires et thermiques, certaines installations sont sujettes à des couplages acousto-mécaniques pouvant nuire fortement à leur bon fonctionnement. La compréhension et la prédiction de ces couplages multi-physiques nécessitent le développement de modèles numériques de très grande précision. Ces modèles sont si coûteux à résoudre qu’il n’est pas envisageable de les utiliser dans des boucles de contrôle ou encore d’optimisation paramétrique. Dans ce manuscrit de thèse, le but est d’exploiter un nombre limité de calculs coûteux pour construire un modèle numérique qui soit de très faible dimension. Ces modèles numériques réduits doivent être capables, en temps réel, de reproduire ces calculs haute-fidélité mais aussi d’extrapoler ces résultats à d’autres points de fonctionnement plus ou moins proches. L’évolution dé petites perturbations compressibles au sein d’un écoulement complexe moyenné est modélisée à partir des équations d’Euler linéarisées dont la nature hyperbolique complique l’application des méthodes de réduction classiques. Les principales problématiques théoriques et numériques qui émergent lors de la construction du système réduit par méthode de projection sont alors exposées. En particulier, les problèmes fondamentaux de la préservation de la stabilité et du contrôle de l’énergie des systèmes réduits sont largement développés et une nouvelle méthode de stabilisation est proposée. Leur sensibilité paramétrique est aussi discutée. Les modèles réduits stables sont ensuite intégrées dans un code de calcul industriel pour prendre en compte des géométries complexes. De plus, la présence de solides dont les parois peuvent être fixes ou mobiles est abordée. En particulier, les petits déplacements de paroi sont modélisés avec une loi de transpiration. Cette condition aux limites est intégrée dans le formalisme du contrôle de façon à lever la difficulté induite par sa non homogénéité. Finalement, les modèles réduits sont exploités pour prédire en temps réel la réponse des systèmes à une loi de contrôle arbitraire. Par exemple, la fréquence et l’amplitude du chargement peuvent varier. Le code de calcul réduit ainsi développé a pour principale vocation de rendre possible des expertises aéroélastiques à faible coût. / In nuclear and thermal power stations, some installations produce acoustics/mechanics coupling which may cause important damage and bad operating performances. Prediction and understanding of these physical phenomena need the development of high-fidelity numerical models which are prohibitive to solve. Therefore, these models cannot be used for control or even parametric optimization applications. In this work, the goal is to use some high-fidelity solutions for building reduced-order models which are able to calculate again these solutions but in real-time, and also to predict solutions for other close configurations. Modelling of compressible disturbances in a complex mean flow is given by hyperbolic linearized Euler equations which create some difficulties to perform classical reduction methods. Theoretical and numerical problems are then introduced when a projection method is applied. In particular, the conservation of stability and the control of energy of reduced-order models are studied and a new stabilization procedure is proposed. Parametric sensitivity is also discussed. Afterwards, stable reduced-order models are developed in an industrial code to consider complex geometries. Furthermore, modelling of solids with fixed or vibrating walls are taken into account. Particularly, small vibrations are modelled thanks to a transpiration law. This boundary condition is implemented in the framework of linear control theory to apply reduction methods. Finally, reduced-order models are tested to predict solutions in real time. For instance, frequency and amplitude of the loading can change. The developed reduced order model should be used for aeroelastic industrial problems with more realistic costs.

In this study the dynamical modeling of the unsteady flow over a flapping wing is considered. The technique is based on collecting instantaneous velocity field data of the flow using Particle Image Velocimetry (PIV), applying image processing to these snapshots to locate the airfoil, filling the airfoil and its surface with proper velocity data, applying Proper Orthogonal Decomposition (POD) to these post-processed images to compute the POD modes and time coefficients, and finally fitting a discrete time state space dynamical model to the trajectories of the time coefficients using subspace system identification (N4SID). The procedure is applied using MATLAB for the data obtained from NACA 0012, SD 7003, elliptic airfoil and flat plate, and the results show that the dynamical model obtained can represent the flow dynamics with acceptable accuracy.

The initial slow viscous growth of the Tollmein-Schlichting wave in a canonical boundary layer transition is absent in bypass and wake-induced transitions. Although there have been a great deal of studies pertaining to bypass transition in boundary layers, the underlying breakdown mechanism is not clearly understood and it continues to be a subject of interest. Similarly, a wake-induced transition caused by Karman wake in the freestream remains poorly understood. The breakdown in this case is caused by anisotropic disturbances containing large scale unsteadiness in the freestream. Differing view points among workers on the transition process have also added to the complexities. In this thesis, bypass and wake-induced boundary layer transitions studied experimentally towards understanding various flow breakdown features are reported. The measurements were made on a flat plate boundary layer in a low-speed wind tunnel. The particle image velocimetry (PIV) technique was extensively used. Various grids were used to generate nearly isotropic freestream turbulence. A circular cylinder was placed at different heights from the plate leading edge to generate Karman wake in the freestream. Two cylinders of different diameters were used to vary the Reynolds number(based on the cylinder diameter). The PIV measurements being simultaneous over a large spatial domain enabled to assess various spatial transitional flow structures. In the case of bypass transition, the streamwise velocity fluctuation, u, is found to exhibit some organized negative and positive fluctuations that dominate the flow during transition, and confirm the simulation results reported in the literature. These positive and negative u fluctuations are found to be associated with the streak unsteadiness. By conditional sampling of these positive and negative u fluctuations, we find that urms (root-mean-squaredof u)can be expressed as a linear combination of urms,f and urms,b,i.e. urms = a(urms,f + urms,b); ais constant, and the subscripts fand bdenote the positive and nega-tive ufluctuations, respectively. Both urms,f and urms,b arefoundto follow the non-modal growth distribution. The wall-normal results clearly show that an inclined shear layer is often associated with an organized structure of negative ufluctuations and an inflectional in-stantaneous velocity profile. These inclined shear layers appear to be similar to those in ribbon-induced transition. The turbulent spot precursor appears to be the vortex shedding from an oscillating in-clined shear layer. Interestingly, the normalized vortex shedding fre-quency is found to be Reynolds number invariant, as in the case of ribbon-induced transition. The present study also confirms the sim-ulated turbulent spot features, including a thin log-law at the break-down stage. The spanwise plane PIV results reveal the signature of streak secondary instability in the flow in terms of symmetric and anti-symmetric streaks oscillations. The initial growth of streak amplitude is followed by a slow decay. The maximum streak amplitude is well above30% of the freestream velocity. These two aspects provide support to the streak instability analysis reported in the literature. While the present wake-induced transition study provides some sup-port to the available numerical simulation and experimental results, some new results have also emerged. The measured sharp rise in the disturbance energy during transition is found to be closer to the simulated result, compared to the difference reported in the literature. The spanwise vortices in the early stage, as also seen in other experimental studies, deform leading to the formation of lambda structures, the signature of which is found by the linear stochastic analysis. With increased Reynolds number and decreased cylinder height from the plate, the physical size of the lambda structure is found to decrease. These lambda structures are often found to appear in a staggered manner in the spanwise plane, as in the case of sub-harmonic boundary layer transition. Although a sub-harmonic peak in the frequency spectra is reported in the literature, as also in the present study, the clear staggered pattern went unnoticed. Streamwise streaks are subsequently generated due to the mean shear stretching of these lambda vortices. The spanwise spacing of these streamwise streaks is found to be comparable with the recent simulation results. Also, these streaks are found to undergo somewhat sinuous-like oscillations, compared to the only varicose type oscillations reported in the literature. The streak amplitude is found to saturate at about 35% of the freestream speed. Here again an inclined shear layer in the wall-normal plane is associated with organized negative u fluctuations and an inflectional instantaneous velocity profile. The movement of the peak urms towards the wall is found to be due to the positive u fluctuation, which follows a hairpin-like structure. The inclined shear layers herein are associated with the lambda or a hairpin-like structure. As in a by-pass transition, an inclined shear layer, vortex shedding from it, the imprint of which is also found in the linear stochastic analysis are present. The normalized high frequency shed vortices is found to be Reynolds number invariant in the present wake-induced transition, as in ribbon-induced and bypass transitions. Compared to the re-cent suggestion that the parent-offspring mechanism is the governing self-sustaining mechanism in the boundary layer, the present study suggests that streak-instability mechanism is also present. The proper orthogonal decomposition(POD) analysis of the experimental data was carried out with an emphasis on the bypass transition case studied. The first few energetic POD modes are found to capture the dominant flow structures, i.e. the organized positive and negative u fluctuations. In the case of bypass transition, the first two energetic POD modes are self-similar, i.e. independent of the freestream turbulent intensity and the Reynolds number. An attempt is also made to construct a low-dimensional model with the POD eigenfunction modes to predict the qualitative dynamics of bypass transition. This has revealed the existence of a traveling disturbance in the bypass transition. On the whole, the present study shows some similar breakdown features in bypass and wake-induced transitions, although more studies in this regard are essential.

Over the past century, great strides have been made in the advancement of mine fire knowledge since the 1909 Cherry Mine Fire Disaster, one of the worst in U.S. history. However, fire hazards remain omnipresent in underground coal mines in the U.S. and around the world. A precise fire numerical analysis (simulation) before any fire events can give a broad view of the emergency scenarios, leading to improved emergency response, and better health and safety outcomes. However, the simulation cost of precise large complex dynamical systems such as fire in underground mines makes practical and even theoretical application challenging. This work details a novel methodology to reduce fire and airflow simulation costs in order to make simulation of complex systems around fire and mine ventilation systems viable. This study will examine the development of a Reduced Order Model (ROM) to predict the flow field of an underground mine geometry using proper orthogonal decomposition (POD) to reduce the airflow simulation cost in a nonlinear model. ROM proves to be an effective tool for approximating several possible solutions near a known solution, resulting in significant time savings over calculating full solutions and suitable for ensemble calculations. In addition, a novel iterative methodology was developed based on the physics of the fluid structure, turbulent kinetic energy (TKE) of the dynamical system, and the vortex dynamics to determine the interface boundary in multiscale (3D-1D) fire simulations of underground space environments. The proposed methodology was demonstrated to be a useful technique for the determination of near and far fire fields, and could be applied across a broad range of flow simulations and mine geometries. Moreover, this research develops a methodology to analyze the tenable limits in a methane fire event in an underground coal mine for bare-faced miners, mine rescue teams, and fire brigade teams in order to improve safety and training of personnel trained to fight fires. The outcomes of this research are specific to mining although the methods outlined might have broader impacts on the other fields such as tunneling and underground spaces technology, HVAC, and fire protection engineering industries. / Ph. D. / With the rapid advancement of technology, the mine fire knowledge has progressed significantly. Atmospheric monitoring and early sensing of heating has improved; the numerical analysis has been expedited with the usage of supercomputers, and more regulations and standards have been set to increase health and safety of miners. In spite of advancements in these areas, fire hazards remain a critical hazard in underground mines. Developing an emergency plan for the safe escape and for fighting the fire is one of the most important issues during a fire event in underground space environments such as mines. A precise fire numerical analysis (simulation) before any fire events can give a broad view of the emergency situation that leads to improving the health and safety issues in the mining industry. Unfortunately, the precise simulation of the large complex dynamical system such as a fire in underground spaces is costly. This work details a cutting edge approach to reduce the fire and airflow simulation costs in order to make simulation of complex systems around fire and mine ventilation systems viable. The main focus of this proposal is to develop novel methodologies to decrease the time of the fire and airflow simulations. The developed methodologies prove to be useful techniques for the reduction of fire simulation and airflow simulation costs. In addition, this study will examine the development of a comprehensive methodology to analyze the tenable limits in a fire event in an underground coal mine in order to improve safety and training of personnel trained to fight fires. These simulations, applied to training, will result in more efficient evacuations (e.g., the decision to leave can be made quickly and with less delay), as well as safe and effective firefighting under certain situations. The target of this research is specific to mining industry although the methods outlined might have broader impacts on the other fields such as tunneling and underground spaces technology, HVAC, and fire protection engineering industries. Therefore, this research may have an immense contribution on the improvement of health and safety associated with firefighting.

La résolution des problèmes de contrôle optimal nécessite des temps de calcul et des capacités de stockage très élevés. Pour s’affranchir de ces contraintes, il est possible d’utiliser les méthodes de réduction de modèles comme la POD (Proper Orthogonal Decomposition). L’inconvénient de cette approche est que la base POD n’est valable que pour des paramètres situés dans un voisinage proche des paramètres pour lesquels elle a été construite. Par conséquent, en contrôle optimal, cette base peut ne pas être représentative de tous les paramètres qui seront proposés par l’algorithme de contrôle. Pour s’affranchir de cet handicap, une méthodologie de contrôle optimal utilisant des modèles réduits adaptatifs a été proposée dans ce manuscrit. Les bases réduites adaptées sont obtenues à l’aide de la méthode d’interpolation ITSGM (Interpolation on Tangent Subspace of Grassman Manifold) ou de la méthode d’enrichissement PGD (Proper Generalized Decomposition). La robustesse de cette approche en termes de précision et de temps de calcul a été démontrée pour le contrôle optimal (basé sur les équations adjointes) des équations 2D de réaction-diffusion et de Burgers. L’approche basée sur l’interpolation ITSGM a également été appliquée avec succès pour contrôler l’écoulement autour d’un cylindre 2D. Deux méthodes de réduction non intrusives, ne nécessitant pas la connaissance des équations du modèle étudié, ont également été proposées. Ces méthodes appelées NIMR (Non Intrusive Model Reduction) et HNIMR (Hyper Non Intrusive Model Reduction) ont été couplées à un algorithme génétique pour résoudre rapidement un problème de contrôle optimal. Le problème du contrôle optimal de l’écoulement autour d’un cylindre 2D a été étudié et les résultats ont montré l’efficacité de cette approche. En effet, l’algorithme génétique couplé avec la méthode HNIMR a permis d’obtenir les solutions avec une bonne précision en moins de 40 secondes. / The numerical resolution of adjoint based optimal control problems requires high computational time and storage capacities. In order to get over these high requirement, it is possible to use model reduction techniques such as POD (Proper Orthogonal Decomposition). The disadvantage of this approach is that the POD basis is valid only for parameters located in a small neighborhood to the parameters for which it was built. Therefore, this basis may not be representative for all parameters in the optimizer’s path eventually suggested by the optimal control loop. To overcome this issue, a reduced optimal control methodology using adaptive reduced order models obtained by the ITSGM (Interpolation on a Tangent Subspace of the Grassman Manifold) method or by the PGD (Proper Generalized Decomposition) method, has been proposed in this work. The robustness of this approach in terms of precision and computation time has been demonstrated for the optimal control (based on adjoint equations) of the 2D reaction-diffusion and Burgers equations. The interpolation method ITSGM has also been validated in the control of flow around a 2D cylinder. In the context of non intrusive model reduction, two non intrusive reduction methods, which do not require knowledge of the equations of the studied model, have also been proposed. These methods called NIMR (Non-Intrusive Model Reduction) and HNIMR (Hyper Non-Intrusive Model Reduction) were developed and then coupled to a genetic algorithm in order to solve an optimal control problem in quasi-real time. The problem of optimal control of the flow around a 2D cylinder has been studied and the results have shown the effectiveness of this approach. Indeed, the genetic algorithm coupled with the HNIMR method allowed to obtain the solutions with a good accuracy in less than 40 seconds.

The spectral-element method (SEM) is used to study wall-bounded turbulent flowsin moderately complex geometries. The first part of the thesis is devoted to simulations of canonical flow cases, such as temporal K-type transitionand turbulent channel flow, to investigate general resolution requirements and computational efficiency of the numerical code nek5000. Large-eddy simulation (LES) is further performed of a plane asymmetric diffuser flow with an opening angle of 8.5 degrees, featuring turbulent flow separation. Good agreement with numerical studies of Herbst (2007) is obtained, and it is concluded that the use of a high-order method is advantageous for flows featuring pressure-induced separation. Moreover, it is shown, both a priori on simpler model problems and a posteriori using the full Navier--Stokes equations, that the numerical instability associated with SEM at high Reynolds numbers is cured either by employing over-integration (dealiasing) or a filter-based stabilisation, thus rendering simulations of moderate to high Reynolds number flows possible. The second part of the thesis is devoted to the first direct numerical simulation (DNS) of a truly three-dimensional, turbulent and separated diffuser flow at Re = 10 000 (based on bulk velocity and inflow-duct height), experimentally investigated by Cherry et al. (2008). The massively parallel capabilities of the spectral-element method are exploited by running the simulations on up to 32 768 processors. Very good agreement with experimental mean flow data is obtained and it is thus shown that well-resolved simulations of complex turbulent flows with high accuracy are possible at realistic Reynolds numberseven in complicated geometries. An explanation for the discovered asymmetry of the mean separated flow is provided and itis demonstrated that a large-scale quasi-periodic motion is present in the diffuser. In addition, a new diagnostic measure, based on the maximum vorticity stretching component in every spatial point, is designed and tested in a number of turbulent and transitional flows. Finally, Koopman mode decomposition is performed of a minimal channel flow and compared to classical proper orthogonal decomposition (POD). / QC 20111206

Afin d’élaborer des stratégies de contrôle des écoulements en temps réel, il est nécessaire d’avoir recours à des modèles d’ordre réduit (ROMs), car la résolution des équations complètes est trop coûteuse en temps de calcul (des jours, des semaines) et en espace mémoire. Dans cette thèse, les modèles réduits ont été construits avec la méthode POD (Proper Orthogonal Decomposition). Une méthode de projection basée sur la minimisation des résidus, initiée par les travaux de Leblond et al. [134] a été proposée. Dans certaines configurations, la précision des résultats est significativement augmentée, par rapport à une projection de Galerkin classique. Dans un second temps, un algorithme d’optimisation non-linéaire, à direction de descente basée sur la méthode des équations adjointes, a été couplé avec des modèles réduits utilisant des bases POD. Deux méthodes de construction de base POD ont été employées : soit avec un paramètre (un nombre de Reynolds,. . . ), soit avec plusieurs paramètres (plusieurs nombres de Reynolds, . . . ). Les ROMs obtenus ont été utilisés pour contrôler la dispersion d’un polluant dans une cavité ventilée puis pour contrôler le champ de température dans une cavité entraînée différentiellement chauffée. Le contrôle est réalisé en temps quasi-réel et les résultats obtenus sont plutôt satisfaisants. Néanmoins, ces méthodes sont encore trop coûteuses en espace mémoire pour être aujourd’hui embarquées dans les boîtiers de contrôle utilisés dans le bâtiment. Une autre stratégie de contrôle, s’appuyant sur les contrôleurs actuels, a ainsi été développée. Celle-ci permet d’obtenir la température (ainsi que la vitesse) dans la zone d’occupation du bâtiment, en utilisant une décomposition des champs par POD et un algorithme d’optimisation de Levenberg-Marquardt. Elle a été validée sur une cavité différentiellement chauffée, puis appliquée sur une cavité ventilée 3D, proche d’un cas réel. / In order to control flows in real-time, it is necessary to resort to reduced-order models (ROMs) because the classical method of simulations is too expensive in CPU time (several days, weeks) and memory storage. In this thesis, the ROMs have been built with the POD (Proper Orthogonal Decomposition) technique. First, a projection method based on the minimization of the equations residuals and established starting from the works of Leblond et al. [134] have been developed. In some cases, the results accuracy is significantly increased. Secondly, a direct descent optimization algorithm based on adjoint-equations has been coupled with POD/ROMs. Two construction methods of POD bases has been employed: either with simulations for only one parameter (one Reynolds number, . . . ), or with simulations for several parameters (several Reynolds numbers,. . . ). The obtained ROMs have been applied in order to control the pollutant dispersion and then to control the temperature field in a lid-driven cavity heated by the left. The control is realized in quasi-real time and the results are rather satisfying. Nevertheless, these methods are still too expensive in memory storage to be embedded in the current controllers. Thus, another control strategy has been proposed, using POD and an optimization algorithm (Levenberg-Marquardt). This one enables to obtain the temperature (and the velocity) in the occupation zone of the building and has been validated on the lid-driven cavity heated by the left and applied on a 3D-ventilated cavity, similar to a real case.

Ce mémoire de thèse est formé de quatre chapitres. Un premier chapitre présente les différentes notions et outils mathématiques utilisés dans le corps de la thèse ainsi qu’une description des résultats principaux que nous avons obtenus. Le second chapitre présente une généralisation d’un résultat obtenu par Rousselet-Chénais en 1990 qui décrit la sensibilité des sous-espaces propres d’opérateurs compacts auto-adjoints. Rousselet-Chénais se sont limités aux sous-espaces propres de dimension 1 et nous avons étendu leur résultat aux dimensions supérieures. Nous avons appliqué nos résultats à la Décomposition par Projection Orthogonale (POD) dans le cas de variation paramétrique, temporelle ou spatiale (Gappy-POD). Le troisième chapitre traite de l’estimation du flot optique avec des énergies quadratiques ou linéaires à l’infini. On montre des résultats mathématiques de convergence de la méthode de Décomposition Progressive Généralisée (PGD) dans le cas des énergies quadratiques. Notre démonstration est basée sur la décomposition de Brézis-Lieb via la convergence presque-partout de la suite gradient PGD. Une étude numérique détaillée est faite sur différents type d’images : sur les équations de transport de scalaire passif, dont le champ de déplacement est solution des équations de Navier-Stokes. Ces équations présentent un défi pour l’estimation du flot optique à cause du faible gradient dans plusieurs régions de l’image. Nous avons appliqué notre méthode aux séquences d’images IRM pour l’estimation du mouvement des organes abdominaux. La PGD a présenté une supériorité à la fois au niveau du temps de calcul (même en 2D) et au niveau de la représentation correcte des mouvements estimés. La diffusion locale des méthodes classiques (Horn & Schunck, par exemple) ralentit leur convergence contrairement à la PGD qui est une méthode plus globale par nature. Le dernier chapitre traite de l’application de la méthode PGD dans le cas d’équations elliptiques variationnelles dont l’énergie présente tous les défis aux méthodes variationnelles classiques : manque de convexité, manque de coercivité et manque du caractère borné de l’énergie. Nous démontrons des résultats de convergence, pour la topologie faible, des suites PGD (lorsqu’elles sont bien définies) vers deux solutions extrémales sur la variété de Nehari. Plusieurs questions mathématiques concernant la PGD restent ouvertes dans ce chapitre. Ces questions font partie de nos perspectives de recherche. / This thesis is formed of four chapters. The first one presents the mathematical notions and tools used in this thesis and gives a description of the main results obtained within. The second chapter presents our generalization of a result obtained by Rousselet-Chenais in 1990 which describes the sensitivity of eigensubspaces for self-adjoint compact operators. Rousselet-Chenais were limited to sensitivity for specific subspaces of dimension 1, we have extended their result to higher dimensions. We applied our results to the Proper Orthogonal Decomposition (POD) in the case of parametric, temporal and spatial variations (Gappy- POD). The third chapter discusses the optical flow estimate with quadratic or linear energies at infinity. Mathematical results of convergence are shown for the method Progressive Generalized Decomposition (PGD) in the case of quadratic energies. Our proof is based on the decomposition of Brézis-lieb via the convergence almost everywhere of the PGD sequence gradients. A detailed numerical study is made on different types of images : on the passive scalar transport equations, whose displacement fields are solutions of the Navier-Stokes equations. These equations present a challenge for optical flow estimates because of the presence of low gradient regions in the image. We applied our method to the MRI image sequences to estimate the movement of the abdominal organs. PGD presented a superiority in both computing time level (even in 2D) and accuracy representation of the estimated motion. The local diffusion of standard methods (Horn Schunck, for example) limits the convergence rate, in contrast to the PGD which is a more global approach by construction. The last chapter deals with the application of PGD method in the case of variational elliptic equations whose energy present all challenges to classical variational methods : lack of convexity, lack of coercivity and lack of boundedness. We prove convergence results for the weak topology, the PGD sequences converge (when they are well defined) to two extremal solutions on the Nehari manifold. Several mathematical questions about PGD remain open in this chapter. These questions are part of our research perspectives.

Bats are some of the most agile flyers in nature. Their wings are highly articulated which affords them very fine control over shape and form. This thesis investigates the flight of Hipposideros Pratti. The flight pattern studied is nominally level and straight. Measured wing kinematics are used to describe the wing motion. It is shown that Proper Orthogonal Decomposition (POD) can be used to effectively to filter the measured kinematics to eliminate outliers which usually manifest as low energy higher POD modes, but which can impact the stability of aerodynamic simulations. Through aerodynamic simulations it is established that the first two modes from the POD analysis recover 62% of the lift, and reflect a drag force instead of thrust, whereas the first three modes recover 77% of the thrust and even more lift than the native kinematics. This demonstrates that mode 2, which features a combination of spanwise twisting (pitching) and chordwise cambering, is critical for the generation of lift, and more so for thrust. Based on these inferences, it is concluded that the first 7 modes are sufficient to represent the full native kinematics. The aerodynamic simulations are conducted using the immersed boundary method on 128 processors. They utilize a grid of 31 million cells and the bat wing is represented by about 50000 surface elements. The movement of the immersed wing surface is defined by piecewise cubic splines that describe the time evolution of each control point on the wing. The major contribution of this work is the decomposition of the native kinematics into canonical flapping wing physical descriptors comprising of the flapping motion, stroke-plane deviation, pitching motion, chordwise, and spanwise cambering. It is shown that the pitching mode harvests a Leading Edge Vortex (LEV) during the upstroke to produce thrust. It also stabilizes the LEV during downstroke, as a result, larger lift and thrust production is observed. Chordwise cambering mode allows the LEV to glide over and cover a large portion of the wing thus contributing to more lift while the spanwise cambering mode mitigates the intensification of LEV during the upstroke by relative rotation of outer part of the wing ( hand wing ) with respect to the inner part of the wing ( arm wing). While this thesis concerns itself with near straight-level flight, the proposed decomposition can be applied to any complex flight maneuver and provide a basis for unified comparison not only over different bat flight regimes but also across other flying insects and birds. / MS / Bats are some of the most agile flyers in nature. Their wings are highly articulated which affords them very fine control over shape and form. This thesis investigates the flight of Hipposideros Pratti. The flight pattern studied is nominally level and straight. Measured wing kinematics are used to describe the wing motion. The central motivation of the thesis is to characterize how the bat uses its wings to generate lift to counter gravity and thrust to move forward against drag forces. A mathematical filter based on Proper Orthogonal Decomposition (POD) is used to filter the measured wing motion to eliminate high frequency noise in the data but at the same time including including the important motions which produce lift and thrust. The filtered native kinematics is decomposed into flapping wing motions comprising of flapping mode, stroke-plane deviation, pitching motion, chordwise, and spanwise cambering. It is shown that the pitching mode harvests the low pressure region created by the Leading Edge Vortex (LEV) during the upstroke to produce thrust. It also stabilizes the LEV during the downstroke, as a result, larger lift and thrust production is observed. Chordwise cambering mode allows the LEV to glide over and cover a large portion of the wing thus contributing to more lift, while the spanwise cambering mode mitigates the intensification of LEV during the upstroke by relative rotation of the outer part of the wing (hand wing) with respect to the inner part of the wing (arm wing). While this thesis concerns itself with near straight-level flight, the proposed decomposition can be applied to any complex flight maneuver and provide a basis for unified comparison not only over different bat flight regimes but also across other flying insects and birds.