Due to their high specific strength and stiffness, honeycomb sandwich structures are used in impact-resistance applications. Their structural efficiency depends to a great extent on the lightweight core separating the face sheets and providing overall high stiffness. Detailed finite element modeling of the penetration of honeycombs by a projectile can be fairly complex, and computationally expensive as shown in the first part of this study. A computationally efficient axisymmetric equivalent numerical homogeneous model for Aluminum 5052-1119 1/8in - 0.001in hexagonal honeycomb subjected to high speed impacts in the range of 60 m/s to 140 m/s is then developed. An equation-of-state model for porous media is used for the equivalent honeycomb medium. A Taguchi optimization, based on four unknown porous material parameters, is carried out. With the optimal set, the equivalent model can accurately predict perforation velocities for different impact conditions. The methodology for the optimization is explained and can be used for any velocity range. The product of this work is a computationally efficient numerical model that requires less than 8% of the time needed to numerically analyze honeycombs in detail.

3D fully viscous turbulent aero-icing simulation remains too computationally intensive when broad range parametric studies are needed, such as during a certification process or anti-icing systems optimization. In order to make such simulations more practical, this work presents a reduced order modeling based on the Proper Orthogonal Decomposition (POD) method which is a popular approach to extract a low-cost CFD approximation from a limited number of "snapshots". POD, equipped with multidimensional Kriging interpolation method, is able to predict the uncalculated solution at intermediate values of three and more input parameters. In this work, this approach is applied to calculate in-flight icing. Also presented are the examples of other current and future applications of POD in aero-icing including droplet impingement calculation and Computational Wind Engineering. The method of selecting the pre-computed CFD solutions (snapshots) and the performance of the Kriging interpolation method are identified as the two main factors which affect the quality of POD approximations. The smoothness of POD coefficients α assumed during Kriging interpolation is examined. / La simulation numérique des écoulements tridimensionnels visqueux turbulents et des phénomènes associés au givrage en vol reste à ce jour trop coûteux en ressources numériques quand il s'agit d'effectuer des études paramétriques, comme lors d'un processus de certification ou de l'optimisation de systèmes d'antigivrage. Dans le but de rendre ce type de simulations plus abordables, une modélisation d'ordre réduit basée sur la méthode de Décomposition Orthogonale Propre (DOP) est présentée dans cette thèse. Cette méthode constitue une approche largement répandue pour extraire une approximation de simulation CFD peu coûteuse à partir d'un nombre limité de solutions préalablement obtenues, appelées "snapshots". La DOP, munie de la méthode d'interpolation multidimensionnelle Kriging, est capable de prédire une solution intermédiaire pour trois paramètres d'entrée ou plus. Cette approche est ici appliquée à la simulation du givrage en vol. D'autres exemples d'applications présentes et futures de DOP dans les domaines du givrage en vol et du Computational Wind Engineering sont aussi présentés. La méthode de sélection des simulations CFD préalablement calculées ("snapshots") ainsi que la performance de la méthode d'interpolation Kriging sont identifiées comme étant les deux facteurs principaux affectant la qualité des approximations obtenues par DOP. La continuité des coefficients α de la DOP utilisés lors de l'interpolation Kriging est examinée.

The present work focuses on the mixing of passive scalars in a turbulent flow (i.e., scalars that do not disturb the velocity field in which they are transported). To this end, small temperature differences are mixed in a turbulent air flow. Of particular interest are (i) the turbulent Prandtl number, and (ii) the mechanical-to-thermal time-scale ratio – two recurring quantities employed by modellers, determined from experiments, and generally assumed to be constant based on the type of flow. The objective of the work is to study the sensitivity of these two quantities to differences in their injection method within the same flow. To achieve this goal, mixed velocity-temperature statistics were measured in the wake of a circular cylinder. The wake was heated by one of two ways: heating the cylinder, or heating an array of fine parallel wires (called a mandoline) placed downstream of the cylinder. The experimental results demonstrate that the magnitude of the turbulent Prandtl number varies throughout the wake for both scalar fields and is consistently greater than the typical value of 0.7 used in turbulence models. Furthermore, the turbulent Prandtl number differs for the two scalar injection methods, despite being within the same flow. Likewise, the values of the mechanical-to-thermal time-scale ratio vary across the wake for both scalar fields and significant differences exist between the two scalar injection methods. Hence, both the turbulent Prandtl number and the mechanical-to-thermal time-scale ratio not only depend on the type of flow, but on the scalar field injection mechanism method as well. This dependence should be taken into account in turbulence models. / Ce travail est une étude du mélange de scalaires passifs en écoulements turbulents. (Un scalaire est dit passif s'il n'exerce pas d'influence sur le champ de vitesse qui le transporte). Par exemple, nous considérons de faibles différences de température mélangées sont considérées au sein d'un écoulement turbulent d'air comme le scalaire passif. Nous sommes particulièrement intéressés par i) le nombre de Prandtl turbulent et ii) le rapport d'échelle des temps dynamique et thermique – deux quantités utilisées pour la modélisation de scalaires passifs en écoulements turbulents. Ces deux quantités sont généralement considérées comme des constantes et déterminées à partir de résultats expérimentaux. L'objectif de cette thèse est d'étudier la sensibilité de ces deux quantités aux différentes méthodes d'injection dans l'écoulement turbulent. Pour atteindre cet objectif, nous réalisons des mesures simultanées de vitesse et de température dans le sillage d'un cylindre. Ce sillage peut être chauffé de deux façons différentes: soit le cylindre lui-même est chauffé, ou une mandoline placée en aval du cylindre est chauffée. Les résultats expérimentaux démontrent que le nombre de Prandtl turbulent varie pour les deux méthodes d'injection du scalaire, bien que le champ dynamique demeure identique dans les deux cas. D'autre part, les valeurs mesurées sont systématiquement plus grandes que la valeur typique de 0.7 utilisée en modélisation. De plus, le rapport d'échelle des temps dynamique et thermique varie lui aussi à travers le sillage pour les deux types d'injection utilisés. Ainsi, nous observons que le nombre de Prandtl turbulent et le rapport d'échelle des temps dynamique et thermique sont à la fois dépendants du champ dynamique et du type d'injection utilisé. Cette dernière dépendance devrait être aussi considérée dans les modèles de turbulence. fr

This thesis undertakes the investigation of the dynamics of two different cases of a slender, flexible pipe conveying fluid: (i) an aspirating cantilevered pipe, ingesting fluid at its free end and transporting it towards the clamped end, and (ii) a pipe conveying a fluid whose density varies axially along the length of the pipe. The general context of the research is first provided by broadly introducing the field of Fluid-Structure Interactions (FSI) and reviewing the basic theory regarding pipes conveying fluid. Subsequently, a numerical approach coupling Computational Fluid Dynamics (CFD) and Computational Structural Mechanics (CSM) to simulate each system is developed in ANSYS™. Lastly, the linear equation of motion is derived for each system and solved using a Galerkin approach; the numerical experiments are then combined with these analytical results to determine the stability characteristics of each system. The problem of an aspirating cantilevered pipe is of both fundamental and practical interest, with applications, for example, in deep sea ocean mining. The motivation for a continued study of the system is demonstrated through a review of previous research on the topic – spanning many years and yielding often contradictory results. The newly proposed analytical model, derived using a Newtonian approach and heavily influenced by CFD analysis, is different from previous ones, most notably because of the inclusion of a two-part fluid depressurization at the intake. In this case, the combined numerical and analytical approaches suggest a first-mode loss of stability by flutter – albeit a very weak one – at comparable but usually lower flow velocities than the discharging cantilever. In the case of a pipe conveying variable density fluid, the analytical model is derived using a Hamiltonian approach, for (i) a pipe clamped at both ends and (ii) a cantilevered pipe. It is shown that these systems lose stability by buckling and flutter respectively, simi / Cette thèse entreprend l'étude de la dynamique de deux types de tuyaux flexibles parcourus par un fluide : (i) un tuyau encastré-libre aspirant le fluide du côté libre et l'amenant vers le côté encastré, et (ii) un tuyau transportant un fluide dont la densité varie axialement au long du tuyau. Le contexte général de cette recherche est d'abord présenté en introduisant le domaine des Interactions fluide-structure (FSI) et en révisant la théorie de base des tuyaux transportant un fluide. Par la suite, une approche numérique couplant la simulation des fluides (CFD) et des structures mécaniques (CSM) est développée dans ANSYS™. Dernièrement, une équation de mouvement linéaire est dérivée pour chaque système et analysée par une méthode Galerkin; les résultats numériques sont enfin combinés avec ces résultats analytiques pour déterminer les caractéristiques de stabilité de chaque système. Le tuyau encastré-libre aspirant est d'un intérêt fondamental et aussi pratique, possédant des applications, entre autres, dans l'industrie minière sous-marine. La raison d'une étude poursuivie est démontrée par un retour sur la recherche antérieure traitant du sujet – s'étendant sur plusieurs années et produisant souvent des résultats contradictoires. Le nouveau model analytique, dérivé utilisant une approche Newtonienne et largement influencé par une analyse CFD, se distingue de ses prédécesseurs notamment par l'inclusion d'une dépressurisation en deux parties à l'entrée. Pour ce cas, les approches numérique et analytique suggèrent tous les deux une perte de stabilité par flottement dans le premier mode – quoiqu'une instabilité très faible – à des vitesses comparables, mais généralement moindre que le cas déchargeant. Dans le cas d'un tuyau transportant un fluide de densité variable, le modèle analytique est dérivé à l'aide d'une approche Hamiltonienne, pour des tuyaux (i) encastré-encastré et (ii) encastr

Travelling long distances with maximum reliability are necessary requirements for future lunar rover missions. Rovers' mobility performance highly depends on wheel type and the mechanical properties of the terrain on which it is rolling. On the lunar surface, the terrain is primarily composed of very fine grained abrasive particles called regolith. Traditional pneumatic rubber wheels are not a viable option for planetary rovers, due to the unknown properties of rubber over a long term exposure to radiation, and the chances of failure in the near vacuum environment. Therefore, non-pneumatic non-rubber compliant wheels have been recognized as a best possible option for planetary exploration rovers.The present research thesis focuses on testing procedures and data analysis of different prototype wheels, rolling on dry sand, by considering several wheel-soil performance parameters including traction, slope climbing ability, rolling resistance, and power consumption. Results from these experiments demonstrate some of the main wheel properties that can affect wheel performance at low speed conditions and provide preliminary data for validation of wheel-terrain model simulation performed within the McGill University research group. The tested wheels were designed and tested at McGill University, Montreal, Canada, as part of a partnership program between the Canadian Space Agency, Neptec Design Group and a number of associated organizations. / Traverser des longues distances avec un maximum de fiabilité sont des conditions nécessaires pour les futures missions des véhicules d'exploration lunaire. La mobilité des "rovers" d'exploration planétaire dépend extrêmement du type de roue et des propriétés mécaniques du terrain sur lequel elle roule. Sur la surface lunaire, le terrain est principalement composé des grains abrasives très fins, appelé "regolith". Les roues pneumatiques traditionnelles ne sont pas une option viable pour les "rovers" d'exploration planétaire, en raison des propriétés inconnues de caoutchouc sur une exposition à long terme au rayonnement, et les hasards d'échec dans le vide. Ce projet de recherche se concentre sur les procédures d'essai et d'analyse des données des différentes roues prototypes roulant sur un terrain déformable, en mesurant plusieurs paramètres de performance y compris la traction, la capacité de monter des pentes, la consommation de pouvoir et la résistance au mouvement de roue sur le sable sec. Les résultats de ces expériences démontrent certaines propriétés de roue qui peuvent affecter la performance de roue roulant à basse vitesse sur le sable sec et fournissent des données préliminaires pour la validation des simulations de modèles de roue sur terrain déformable qui ont été faites parmi le groupe de recherche à l'université McGill Montréal, Canada. Les roues prototypes ont été conçues et testées à l'université McGill, dans le cadre d'un programme de partenariat entre l'Agence Spatiale Canadienne, Neptec Design Group et un nombre d'organisations associées.

The Gurney flap can provide an increase in lift with a small drag penalty up to a flap height of approximately 3% of the chord. For flap height larger than 3% chord (herein referred to as Gurney-type flap), there is a significant increase in drag, causing a deteriorating lift-to-drag ratio. In this work, perforation was introduced in large Gurney-type flaps in an attempt to improve their aerodynamic performance. A study of different flap heights and perforation porosities were performed by using force balance, surface pressure and hot-wire measurements in conjunction with particle image velocimetry (PIV). Results show that flap porosity reduced lift in comparison to the solid flap due to the decrease in positive camber effects and decompression of the cavity flow up stream of the flap. The corresponding reduction in drag, however, outweighed the loss in lift and rendered an improved lift-to-drag ratio. Detailed PIV flow field behind the perforated flap revealed that the existence of perforation-generated jets are responsible for the observed differences in aerodynamic performance. If strong enough, perforation- generated jets could eliminate the vortex shedding process behind the flap. Furthermore, the near wake was found to be disrupted and narrowed, drastically suppressing fluctuation intensity. / Le volet de Gurney peut augmenter la force de sustentation (portance) avec peu de pénalité sur la traînée aérodynamique jusqu'à un hauteur de volet d'environ 3% de la corde du profil. Pour les volets plus hauts que 3% de la corde (appelles "Gurney-type flaps" ici), il y a une croissance significative à la traînée aérodynamique, causant une détérioration de la finesse (rapport portance/traînée). Dans cette étude, la perforation a été introduite dans les grands volets de Gurney afin d'essayer d'améliorer la finesse. Différentes hauteurs de volet et porosités de perforation ont été étudiées par balance de force, pression superficielle, fil chaud, et vélocimétrie par image de particules (PIV). Les résultats démontrent que la perforation des volets réduit la portance comparés aux cas non-perforés à cause de la diminution d'effet de cambrure positive et de la décompression de l'écoulement de la cavité en amont du volet. La réduction correspondante de la traînée, cependant, était supérieure à la perte de la portance et a amélioré la finesse. Les champs d'écoulement détaillés obtenus par PIV en aval du volet ont également indiqué que l'existence de jets d'air générés par la perforation sont responsable des différences observées à la finesse. Les jets générés par la perforation, si assez forts, peuvent éliminer le processus de décollement de tourbillon en aval du volet. En outre, le sillage proche s'est avéré perturbé et rétréci, supprimant l'intensité de fluctuation.

The use of carbon fibre reinforced composites for aerospace structures has seen a high increase in recent years, and is still growing. The high stiffness-to-weight ratio of these materials makes them ideal for primary structures on airplanes, satellites, and spacecrafts. Nevertheless, the manufacturing of composites remains very costly since it requires equipment investment such as an autoclave, and very qualified workers. Out-of-autoclave manufacturing technology is very promising since it only requires a traditional oven, while still aiming at similar part quality. However, the absence of positive pressure compared with an autoclave makes it more difficult to achieve low porosity parts. This research investigates the manufacturing of complex features with out-of autoclave prepreg technology. The features studied are tight-radius corners with a curvature change, and ply drop-offs. Ply drop-offs tests were conducted to identify if porosity is higher at ply terminations. In corners, the bagging arrangement was modified to achieve the most uniform thickness in areas of curvature change, even with small radii. The conclusions from these studies provided us with guidelines to manufacture larger representative parts, which included these features. The representative parts were tested for porosity, thickness uniformity, mechanical performance, and glass transition temperature(Tg). A total of four representative parts were manufactured with out-of-autoclave technology, and one more was manufactured with an autoclave to allow for a proper comparison between the two processes. The materials used were MTM45-1 5 harness satin and CYCOM5320 plain weave for the out-of-autoclave parts,and CYCOM5276-1 plain weave for the autoclave part. The effect of ply dropoffs on porosity was found to be negligible. Thickness deviation in corners was attributed to a combination of consumable bridging, prepreg's bulk factor and inter-ply shear. Overall, out-of-autoclave prepregs showed performance similar to autoclave prepregs. / L'utilisation de matériaux composites en fibres de carbone pour des structures aéronautiques a connu une croissance rapide ces dernières années, et continue de croitre. Le rapport raideur/masse de ce type de matériaux en fait une solution idéale pour les structures primaires d'avions, de satellites, ou de navettes spatiales. Toutefois, la fabrication de ces pièces en composites demeure extrêmement couteuse puisqu'elle nécessite de lourds investissements d'équipement tels que l'acquisition d'un autoclave, ainsi que de la main-d'oeuvre qualifiée. La technologie hors autoclave semble très prometteuse puisqu'elle ne requiert que l'utilisation d'un four traditionnel, tout en visant à obtenir des pièces de qualité similaire. Cependant, l'absence de pression extérieure provenant de l'autoclave rend plus délicate l'obtention de pièces ayant une faible porosité. Cette recherche a pour thème la fabrication d'éléments complexes avec la technologie hors autoclave. Les éléments étudiés sont des angles convexes et concaves ayant de faibles rayons de courbure, ainsi que des plis partiels. Des tests sur les plis partiels ont été réalisés pour déterminer si ils sont associés à une augmentation de la porosité. Dans les angles, l'arrangement des consommables a été modifié pour obtenir l'épaisseur la plus uniforme possible dans les zones de changement de courbure, et cela même pour de faibles rayons. Les conclusions de ces tests nous ont permis de considérer la fabrication de pièces représentatives de plus grande taille, et qui contiennent les éléments précédemment étudiés. Les pièces représentatives ont été testées pour déterminer leur niveau de porosité, l'uniformité de leur épaisseur, leur performance mécanique, et leur température de transition vitreuse. Au total, quatre pièces représentatives ont été fabriquées par technologie hors autoclave, et une a été fabriquée dans un autoclave afin de permettre une comparaison de bon aloi entre ces deux procédés de fabrication. Les matériaux utilisés pour cette recherche étaient du MTM45-1 5 harness satin et du CYCOM5320 plain weave pour les pièces hors autoclave, ainsi que du CYCOM5276-1 plain weave pour la pièce autoclave. La présence de plis partiels n'a pas été associable à une augmentation notable de la porosité. L'uniformité d'épaisseur s'est révélée être une combinaison de pontage des consommables, du facteur de foisonnement du pré-imprégné, et du cisaillement entre les plis de fibre. Globalement, les pré-imprégnés hors autoclave ont montré des performances similaires aux pré-imprégnés autoclave.

Experiments were conducted to study oscillation and friction effects on the mass exchange between the recirculating flow in a square basin and the main flow in an open channel. During the experiments, a turbulent mixing layer was observed to form between the recirculating flow in the basin and the main flow. Dye was introduced into the turbulent mixing layer as a tracer to observe the motion of the eddies. The concentration of the dye was measured using a video imaging method. The data were analysed for an early period shortly after the dye entered the basin, and a late period when the dye in the basin had reached a quasi-steady state. Parameters such as retention time, accumulation time and arrival time were estimated using the time series of the concentration in the core of the basin and in the region outside of the core. The circulation time and circulation flow rate were computed using the area occupied by the dye. The rate of increase in the area was equated to the circulation flow rate. The mean and root-mean-square dye-concentration profiles were obtained for a number of cross sections along the mixing layer. The flow in the mixing layer was observed to oscillate due to feedback from the impingement of the mixing layer on the downstream corner of the basin. The oscillation frequencies were determined by spectral analysis of the concentration data at a location in the mixing layer near the point of the impingement. / Tests were conducted for a range of Froude number varying from Fr = 0.09 to 0.47 and a range of bed-friction number from S = 0.023 to 0.27. Self-sustaining oscillations of the type known as "fluid-resonant feedback" were observed in the mixing layer. The onset of the oscillations in the mixing layer was detected to occur for a Froude number greater than 0.28. However, the mixing layer was not responsive to the feedback when the Froude number was less than 0.28. The dimensionless frequency of the oscillation was correlated with the Froude number. The Strouhal and Froude number relation analogously was consistent with the Rossiter's formula, which was developed for cavity oscillations in compressible flow. Beside the flow oscillations, the mixing layer and the recirculating flow in the basin were affected by the friction effect. Friction slow down the circulation in the basin. It changes the turbulence in the mixing layer. These effects of the friction were only significant when the water depth is sufficiently shallow. The dimensionless parameter for the friction effects was the bed-friction number. The demarcation of the flow between the deep and the shallow was found to occur at a bed-friction number S ≃ 0.09. The friction effect was negligible when the bed-friction number is less than the critical value of 0.09.

A computational methodology for the optimization of the thermal performance of straight ducts with non-twisted, uninterrupted, longitudinal internal fins is formulated and demonstrated in the context of steady, laminar, fully-developed forced convection, accounting for conjugate heat conduction in the fins. The fluid flow and heat transfer problems are solved using control-volume-based finite difference and finite element methods. The optimization is done using a gradient method. The fins shapes are approximated using non-uniform rational B-splines (NURBS). The control points of the NURBS curves are among the design variables. In the demonstration problems, the objective was to maximize the total rate of heat transfer to the fluid per unit length of the duct, subject to the constraint of keeping the corresponding pumping power constant. Results pertaining to four sample ducts with fins made of stainless steel, aluminum, and copper, and air as the working fluid, are presented and discussed.

Springback is a phenomenon that occurs when nonhomogeneous elastic and plastic deformation occurs throughout a component during forming processes. Since the elastic deformation is recovered when the forming load is removed, geometrical changes occur. Springback is particularly important in sheet forming; therefore, in order to provide formed parts of close tolerances it is essential to first have a good understanding of the factors which affect springback and to be able to determine the extent those factors have. It is also important to be able to predict springback under different conditions so that it can either be minimized or properly accounted for in the process design. This research presents work done thus far on understanding and predicting springback in sheet metal forming of stainless steel 410 and inconel 718, which have a wide range of usage in the aerospace industry as well as dual-phase steels 6001300 and 600/400 which are used in the automotive sector. / The role that the anisotropy plays in the springback is assessed in this work. The effect of normal anisotropy on springback for the aerospace materials is considered while the automotive materials were considered perfectly isotropic and only the effect of forming conditions was studied. In order to characterize the materials and their anisotropic behaviour, a series of mechanical tests is conducted. These tests include standard uniaxial tension and uniaxial tension-compression tests. Moreover, a series of simple multiplebending experiments were conducted on the aerospace materials (steel and Ni based alloys) to examine the effect of specimen orientation on the springback in simple bending. / Furthermore, since process conditions have an obvious effect on springback and one of the most important of these is the blank holding force, 2D draw bending experiments are conducted with varying blank holding force to assess its role in springback of the formed part. The combined effect of anisotropy and blank holding force was also studied for the aerospace materials. / Finite element simulations that include only either classical isotropic hardening or kinematic hardening did not show close agreement with the experimental findings especially for springback prediction. Therefore, to properly simulate springback a material model that combines both hardening effects, along with the material anisotropy, has been developed in this work. The developed finite element model implements isotropic hardening as well as kinematic hardening based on the Mroz multiple-yield surface formulation. Hill's 1948 yield function with normal anisotropy is considered. The developed material model has been tested by simulating the tension-compression experiments and a good agreement was reached. / Furthermore, to demonstrate the model capability, bending experiments were simulated. Springback angles predicted by the model reflected those obtained experimentally for the simple multiple-bending experiments. Moreover, draw-bending experiments were simulated with the developed material model, which showed good agreement with the experiments. / Finally, the capability of the model can be readily extended to cover real forming operations, which will reduce cost and enhance the quality of the formed parts.