The electromagnetic machines like the dc, induction, synchronous motor/generator and the transformer have an energy flow framework that is similar. All these machines deal with electrical energy in the electrical domain that is interfaced with the magnetic domain. Except for the transformer, the other machines also have one more energy interface i.e. with the mechanical domain. In all these machines, the magnetic domain acts as the silent energy manager. The electrical and the mechanical domain energies will have to pass through the magnetic domain and appropriately get routed. In recognition of the commonality of this pattern of energy flow, this thesis proposes a generalised model of a multiphase electromagnetic device wherein the dc machine, induction machine, synchronous machine and the transformers are special cases of the proposed generalised model. This is derived using bond graphs that is based on the underlining principle of Energy Flow rooted in the concept of Conservation of Energy. A model is a set of mathematical equations representing a physical system. A model is as good as a modeller understanding of the physical system and the underlying approximation he makes while writing down the equations describing the models behaviour to the stimulus. A modelling language tool, which can cut down the approximations made by using the power of identified analogous characteristics across the physical domain, can help make a model more close to real life situation. Bond Graph is such a modelling language which is powerful enough to model the non-linear, multi-disciplinary, hybrid continuous-discrete phenomena encountered in a real life physical system. Bond graphs as a modelling tool was introduced by Professor H.Paynter at Massachusetts Institute of Technology in 1959. The Bond Graph methodology is based on consideration of energy flows between the ports of the components of an engineering system. Bond Graph methodology enables one to develop a graphical model that is consistent with the first principle of energy conservation without having the need to start with establishing and reformulating equations. The derivation of a mathematical model from the graphical description is automated by software tools. As a consequence, a modeller using this methodology can focus on modelling of the physical system. In the graphical representation of bond graph the vertices of a bond graph denote subsystems, system components or elements, while the edges, called power bonds, represent energy flows between them. The nodes of a bond graph have power ports where energy can enter or exit. Bond graph models are developed in a hierarchical top-down or bottom-up approach by using component models or elements from model libraries. An electromagnetic machine is a black box having an assemblage of windings in iron resulting in a combination of input/output ports on shaft and electrical terminals. Abstraction of an machine model by a modeller matching the vision of the observer above is an ideal goal. Bond graph methodology is an appropriate tool for trying to reach this goal as it is based on object oriented modelling techniques. There have been few attempts to model electric machine in bond graph earlier. A well established DC motor bond graph has been widely used in all bond graph literature. But AC rotating machine being a higher order nonlinear system poses a tougher challenge. Here too, there have been few attempts in modelling AC machines. It is observed that majority of AC machine bond graph models have been built up from their mathematical models. But as the bond graph modelling technique is based on the unifying theory of energy exchange, better insight into the system is achievable if the model is conceptualised from its physical structure. This thesis starts from the basic theory of energy port to conceptualise the generalised model from physical correspondence. In this thesis a Rotating Electrical Machine is studied as a physical system. The energy ports inside this physical system is identified. When a physical system receives the energy through its energy port in one energy cycle, it processes this energy in one of the three ways. The received energy is converted into useful work or it is dissipated or stored. The storage can further be classified into two ways, either as kinetic energy or as potential energy. For a rotating electric machine the input-output port for energy exchange are either in electrical or mechanical domain depending on the class of the machine. The magnetic domain across all class of electromagnetic device acts as the energy manager. In order to capture the features of the energy jumping across the air gap in a rotating electrical machine, wherein the magnetic fields from spatially distributed windings of the stator and rotor interplay, an Axis Rotator (AR) element -a mathematical commutator, is introduced in this thesis as a new bond graph element. In a multiphase device, the energy from the various phases and spatial axes are transferred through the axis rotator element. The Axis Rotator is a critical element which helps distinguish between the various classes of electromagnetic devices. The defining features of the Axis Rotator helps in deriving the various special electromagnetic devices (such as the dc machine, induction machine, synchronous machine and the transformer) from the generalised model. The Axis Rotator exists in the magnetic domain. It naturally inherits the characteristics of the magnetic domain. The Axis Rotator as a bond graph element is complex. In a specific case of 3φ Induction Motor an alternative bond graph model with all integral elements is developed. By one to one correspondence with the AR bond graph model, the inner component of ’AR’ can be identified. Another advantage of using this model is that saturable and non-saturable magnetic permeance can be separated out, a useful feature in the nonlinear model discussed next. One of the most distinguishing features of the magnetic domain is the existence of Magnetic Hysteresis. Magnetic Hysteresis is a well understood and studied subject. But this physical process is wilfully ignored by the modelling community at large. The main reason for this is the difficulty of modelling a nonlinear phenomena. The bond graph modelling naturally allows the inclusion of such non-idealities within its framework. This thesis proposes the generalised model along with the inclusion of magnetic non-linearities and non-idealities into the model of the system. This inherent strength of bond graph model flows from the fact that the models in bond graph are developed from the first principles of energy conversation and the mathematical equations are derived later from the evolved graph. The tools that are available for bond graph simulation are not adequate for power electronics systems. The existing tools do not address space vectors and frame transformations. As a consequence it is difficult to simulate the electromagnetic device models developed in this thesis. The need for a bond graph tool to address vectors and frame transformations, a common occurrence in electric machines dynamic model study was acutely felt. This necessitated a support for handling complex data class from the underlying mathematical engine of the software. MATLAB/Simulink is the commonly available mathematical tool which has a support for complex variables. Therefore during the course of this research work a new software tool box was developed which meets the need of electromagnetic machines in particular and other engineering domains in general. For developing the new bond graph simulation software, the language extender approach was chosen, as it combines the capabilities of existing popular mathematical engine with its tested graphical frontend and the flexibility of combining different modelling technique like bond graph, block diagram, equations etc. It also ensures portability as they are compiled by interpreted language compiler of the mathematical engine and are thus independent of the computer operating system. C-MEX S-function methodology was used to develop the software as it has access to lower level functions and methods of the underlying mathematical engine. This helps in speeding up the software execution time alongwith the flexibility in defining new complex elements like the Nonlinear Axis Rotator. In conclusion, this thesis makes the following contributions: (i) The Axis rotator concept to handle space vectors and frame transformations, (ii) generalised model of the electromagnetic device, (iii) introduction of the saturation and hysteresis non-linearity in the magnetic domain, (iv) development of the bond graph toolbox to handle vector and frame transformations.

Les structures soumises à des vibrations sont rencontrées dans diverses applications. Dans denombreux cas, elles sont de nature linéaires, mais quand les amplitudes des oscillations deviennentimportantes, cela provoque un comportement non-linéaire. Par ailleurs, les oscillations desstructures dans un milieu fluide entrainent une interaction fluide-structure. Cette thèse porte surla modélisation du problème fluide-structure non-linéaire. Les cas de non-linéarités étudiés sont lanon-linéarité grands-déplacements caractéristique des structures minces, la non-linéarité localiséegéométrique décrivant une liaison non-linéaire entre deux structures et la non-linéarité acoustiqueparticularité des très hauts niveaux de pression.Pour la modélisation de ces problèmes, il se peut que le calcul en réponse demeure infaisable enraison du temps de calcul. D’une part, on est amené à résoudre des systèmes matriciels (symétriquesou non) de grandes tailles générés par la méthode des éléments finis et d’autre part, cetterésolution demande une évaluation de la force non-linéaire à chaque itération. Afin de diminuer lecoût de calcul, la réduction de modèle par des bases de réductions couplées avec un algorithmeparallélisant l’évaluation de la force non-linéaire, est une alternative à la résolution du systèmecomplet. La construction des bases de réduction doit s’adapter au mieux à chaque problème traité.La base modale du problème linéaire est une première approximation puis elle est enrichie par desinformations qui proviennent à la fois de la nature du couplage et du comportement non-linéaire / Structures subjected to vibrations are found in various applications. In many cases, they behave ina linear way, but when the amplitudes of the oscillations become important, it causes a nonlinearbehavior. Moreover, the oscillations of structures in a fluid field lead to a fluid-structureinteraction. This thesis focuses on the modeling of nonlinear fluid-structure problem. Differentkind of nonlinearities are studied in this work including the large-displacement nonlinearitycharacteristic of thin structures, the localized geometrical nonlinearity describing a nonlinear linkbetween two structures, and the acoustic nonlinearity characteristic of very high levels ofpressure.Modeling such problems are time and memory consuming, that may lead to a limitations of themodel. Therefore, it is necessary to solve a large matrix system (either symmetric or not)generated by the finite element method and the resolution needs an evaluation of the nonlinearforce at each iteration. In order to reduce the computational cost, model reduction with reducedbases combined with parallelization of the nonlinear force evolution is proposed as an alternative tothe resolution of complete systems. Building reduction bases must be adapted to each concernedproblem. The eigenmode of the linear problem is a first approximation and it is enriched withinformation coming from both coupling and nonlinear behaviors.

Le travail de thèse porte sur de nouveaux compléments et améliorations pour la théorie de la péridynamique concernant la modélisation numérique de structures minces telles que les poutres et les plaques, les composites isotropes et multicouches soumis à un chargement dynamique. Nos développements ont principalement porté sur l'exploration des possibilités offertes par la méthode péridynamique, largement appliquée dans divers domaines de l'ingénierie où des discontinuités fortes ou faibles peuvent se produire, telles que des fissures. La procédure de généralisation de la méthode Peridynamics pour la modélisation des structures de poutres de Timoshenko et des structures de plaques de Reissner-Mindlin avec une large plage de rapport épaisseur sur longueur allant de structures épaisses à très minces est indiquée. Et un impact avec une faible vitesse simplifié basé sur le modèle péridynamique développé pour la poutre de Timoshenko et la plaque de Reissner-Mindlin a été proposé en utilisant une procédure de contact spécifique pour l'estimation « naturelle » de la charge d'impact. L’originalité de la méthode actuelle réside dans l’introduction avec deux techniques permettant de réduire le problème de blocage par cisaillement qui se pose dans les structures à poutres et à plaques minces, à savoir la méthode d’intégration réduite (ou sélective) et la formulation mixte. Le modèle péridynamique résultant pour les structures de poutre de Timoshenko et les structures de plaque de Reissner-Mindlin est efficace et ne souffre d'aucun phénomène de verrouillage par cisaillement. En outre, la procédure de généralisation de la méthode péridynamique pour la modélisation de structures composites minces renforcées par des fibres est introduite. L’approche péridynamique pour la modélisation d’une couche est d’abord validée en quasi-statique, ce qui inclut des problèmes de prévision de la propagation de fissures soumis à des conditions de chargement mécaniques. La méthode péridynamique a ensuite été étendue à l’analyse de structures composites minces renforcées par des fibres utilisant la théorie fondamentale d’une couche. Enfin, plusieurs applications impliquant des structures composites minces renforcées par des fibres et des résultats numériques ont été validées par comparaison à la solution FEM obtenue à l'aide d'un logiciel commercial ou à des solutions de référence de la littérature. Dans toutes les applications, Péridynamics montre que les résultats correspondent parfaitement aux solutions de référence, ce qui prouve son potentiel d’efficacité, en particulier pour la simulation de chemins de fissures dans les structures isotropes et composites. / This thesis introduces some new complements and improvments for the Bond-Based Peridynamics theory concerning the numerical modeling of thin structures such as beams and plates, isotropic and multilayer composites subjected to dynamic loading. Our developments have been focused mainly on exploring the possibilities offered by the Peridynamic method, which has been widely applied in various engineering domains where strong or weak discontinuities may occur such as cracks or heterogeneous media. The generalization procedure of the Peridynamics method for the modeling of Timoshenko beam structures and Reissner-Mindlin plate structures respectively with a wide range of thickness to length ratio starting from thick structures to very thin structures is given. And A simplified low velocity impact based on the developed Peridynamic model for Timoshenko beam and ReissnerMindlin plate has been proposed by using a specific contact procedure for the estimation of the impact load. The originality of the present method was the introduction for the first time of two techniques for the alleviation of the shear locking problem which arises in thin beam and plate structures, namely the reduced (or selective) integration method and mixed formulation. The resulting Peridynamic model for Timoshenko beam structures and Reissner-Mindlin plate structures is efficient and does not suffer from any shear locking phenomenon. Besides, the generalization procedure of Peridynamic method for the modeling of fiber-reinforced thin composite structures is introduced. The Peridynamic approach for the modeling of a lamina is firstly validated in the quasi-statics including a crack propagation prediction problems subjected to mechanical loading conditions and then the Peridynamic method was further extended to analyze fiber-reinforced thin composite structures using the fundamental lamina theory. Finally, several applications involving fiber-reinforced thin composite structures and numerical results were validated by comparison to the FEM solution obtained using commercial software or to reference solutions from the literature. In all applications, the Peridynamics shows that results are matching perfectly the reference solutions, which proves its efficiency potentiality especially for crack paths simulation in isotropic and composite structures.

En raison d’un désordre aléatoire à longue distance, un verre présente une symétrie d'inversion telle que la génération de seconde harmonique (GSH) est interdite. Cependant, par irradiation avec un laser femtoseconde (fs) très focalisé, il est possible de précipiter des cristaux optiquement non linéaires, et de rompre la symétrie d'inversion et donc d'induire une GSH. De plus, ceci peut être réalisé localement en trois dimensions. Pour la démonstration, on a appliqué, dans le système vitreux Li₂O-Nb₂O₅-SiO₂ le mode opératoire décrit ci-dessous qui permet la formation de cristaux de LiNbO₃, hautement optiquement non linéaire. La procédure est la suivante : 1) ajustement de la composition chimique du verre pour obtenir un verre suffisamment sensible au laser fs ; 2) contrôle des paramètres du laser (durée d'impulsion, fréquence de répétition des impulsions, vitesse de balayage du faisceau, énergie d'impulsion…) pour obtenir des nanocristaux avec répartition spatiale et taille correcte. En outre, la taille de la zone affectée doit être limitée ; 3) contrôle de l'orientation des nanocristaux. On montre qu'il est possible de satisfaire à cette condition, en contrôlant l'orientation de la polarisation du laser. Ceci a été montré par la méthode de rétrodiffusion d'électrons de diffraction (EBSD). En d'autres termes, ce processus peut être contrôlé directement avec la lumière. En outre, la spectroscopie par rayons X à dispersion d'énergie couplée à la microscopie à transmission électronique à balayage (STEM /EDS) et la microscopie électronique à transmission a révélé une microstructure orientable similaire à celle appelée nanoréseaux formée dans silice. L'originalité est que les nanocristaux optiques non linéaires texturées noyées dans un réseau de "murs" vitreux, sont alignés perpendiculairement à la direction de polarisation du laser. Il en résulte que la biréfringence et la propriété optique non linéaire peuvent être maîtrisées ensemble. Ceci est une percée dans ce travail de thèse. Ces résultats mettent en évidence des modifications spectaculaires de verre par rayonnement laser fs. Avec de nouvelles améliorations dans les techniques de fabrication, l'application de ce travail est de parvenir à réaliser un guide d'ondes biréfringent doubleur ou changeur de fréquences. / Due to random disorder, a glass exhibits inversion symmetry such that second harmonic generation (SHG) is forbidden. However, by irradiation with a tightly focused femtosecond (fs) laser, it is possible to induce nonlinear optical crystal precipitation, in order to break the inversion symmetry and thus to induce SHG. Moreover, this can be achieved locally in three dimensions. For demonstration, we applied the procedure described below in the glass system Li₂O-Nb₂O₅-SiO₂ that allows the formation of LiNbO₃ crystal, a highly non linear optical one. The procedure is thus the following: 1) adjustment of the glass chemical composition for obtaining a glass sensitive enough to fs laser. 2) control of the laser parameters (pulse duration, pulse repetition rate, speed of beam scanning, pulse energy…) for obtaining nanocrystals with correct space distribution and size. In addition, the size of the affected zone has to be limited. 3) control of the orientation of the nanocrystals. We show that it is possible to fulfill this condition by controlling the laser polarization orientation. This has been achieved by electron backscatter diffraction method (EBSD). In other words, this process can be controlled with light directly. In addition, energy dispersive X-ray spectroscopy coupled to scanning transmission electron microscopy (STEM/EDS) and transmission electron microscopy revealed an orientable microstructure similar to the one called nanogratings form in silica. The originality here is a textured nonlinear optical nanocrystals embedded in a network of “walls” made of vitreous phase, aligned perpendicular to the laser polarization direction. It results that birefringence and nonlinear optical property can be mastered in the same time. This is a highly valuable aspect of the work. These findings highlight spectacular modifications of glass by fs laser radiation. With further improvements in the fabrication techniques, the application of this work is to achieve SHG waveguide and birefringence-based devices.

No description available.

Die vorliegende Arbeit beschäftigt sich mit Mechanismen, die in Einzelzellmodellen zu einer frequenzabhängigen Informationsübertragung führen können. Um dies zu untersuchen, werden Methoden aus der theoretischen Physik (Statistische Physik) und der Informationstheorie angewandt. Die Informationsfilterung in mehreren stochastischen Neuronmodellen, in denen unterschiedliche Mechanismen zur Informationsfilterung führen können, werden numerisch und, falls möglich, analytisch untersucht. Die Bandbreite der betrachteten Modelle erstreckt sich von reduzierten strombasierten ’Integrate-and-Fire’ (IF) Modellen bis zu biophysikalisch realistischeren leitfähigkeitsbasierten Modellen. Anhand numerischer Untersuchungen wird aufgezeigt, dass viele Varianten der IF-Neuronenmodelle vorzugsweise Information über langsame Anteile eines zeitabhängigen Eingangssignals übertragen. Der einfachste Vertreter der oben genannten Klasse der IF-Neuronmodelle wird dahingehend erweitert, dass ein Konzept von neuronalem ’Gedächtnis’, vermittelst positiver Korrelationen zwischen benachbarten Intervallen aufeinander- folgender Spikes, integriert wird. Dieses Model erlaubt eine analytische störungstheoretische Untersuchung der Auswirkungen positiver Korrelationen auf die Informationsfilterung. Um zu untersuchen, wie sich sogenannte ’unterschwelligen Resonanzen’ auf die Signalübertragung auswirken, werden Neuronenmodelle mit verschiedenen Nichtlinearitäten anhand numerischer Computersimulationen analysiert. Abschließend wird die Signalübertragung in einem neuronalen Kaskadensystem, bestehend aus linearen und nichtlinearen Elementen, betrachtet. Neuronale Nichtlinearitäten bewirken eine gegenläufige Abhängigkeit (engl. "trade-off") zwischen qualitativer, d.h. frequenzselektiver, und quantitativer Informations-übertragung, welche in allen von mir untersuchten Modellen diskutiert wird. Diese Arbeit hebt die Gewichtigkeit von Nichtlinearitäten in der neuronalen Informationsfilterung hervor. / Neurons transmit information about time-dependent input signals via highly non-linear responses, so-called action potentials or spikes. This type of information transmission can be frequency-dependent and allows for preferences for certain stimulus components. A single neuron can transmit either slow components (low pass filter), fast components (high pass filter), or intermediate components (band pass filter) of a time-dependent input signal. Using methods developed in theoretical physics (statistical physics) within the framework of information theory, in this thesis, cell-intrinsic mechanisms are being investigated that can lead to frequency selectivity on the level of information transmission. Various stochastic single neuron models are examined numerically and, if tractable analytically. Ranging from simple spiking models to complex conductance-based models with and without nonlinearities, these models include integrator as well as resonator dynamics. First, spectral information filtering characteristics of different types of stochastic current-based integrator neuron models are being studied. Subsequently, the simple deterministic PIF model is being extended with a stochastic spiking rule, leading to positive correlations between successive interspike intervals (ISIs). Thereafter, models are being examined which show subthreshold resonances (so-called resonator models) and their effects on the spectral information filtering characteristics are being investigated. Finally, the spectral information filtering properties of stochastic linearnonlinear cascade neuron models are being researched by employing different static nonlinearities (SNLs). The trade-off between frequency-dependent signal transmission and the total amount of transmitted information will be demonstrated in all models and constitutes a direct consequence of the nonlinear formulation of the models.

The aim of the current dissertation is the development of finite element models capable of predicting the damping and the damped structural dynamic response of laminated composite blades and beams. The present thesis is divided into two main parts, of which the first one studies the material coupling effect on the static and modal characteristics of composite structures. New damping coupling terms are formulated and incorporated into a linear beam finite element to better capture the composite material and structural coupling effects. The second part describes the theoretical framework for predicting the nonlinear damping and damped vibration of laminated composite structures due to large in-plane tensile and compressive forces. A nonlinear beam finite element for composite strips is developed capable of capturing the effects of geometric nonlinearity on the damping of composite laminates. The damping mechanics consider a strain based Kelvin viscoelastic model and Green-Lagrange nonlinear strain expressions, which introduce geometric nonlinearity into the formulation. Incorporation of first-order shear deformation theory into the equations of motion provides the linear and new nonlinear cross-section stiffness and damping terms. Within each element, the stain field is approximated by linear interpolation shape functions. An incremental-iterative methodology is formulated into the finite element solver, based on the Newton-Raphson technique in order to obtain the system solution at each iteration, till the final convergence is achieved. For the sake of completeness, a series of experimental measurements were carried out for the composite strip, subject to tensile and buckling loads. Correlations with theoretical predictions gave credence to the ability of the nonlinear finite element to predict damping of composite structures undergoing large displacements and rotations in the nonlinear regime. The finite element was further extended to include the nonlinear analysis of large-scale hollow composite structures. New first- and second-order stiffness and damping terms were developed and incorporated into an updated nonlinear beam finite element, capable of capturing the effect of rotational stresses on the static and modal characteristics of composite beams and blades. / Σκοπός της παρούσας διδακτορικής διατριβής με τίτλο: "Ανάπτυξη Μη-Γραμμικού Προτύπου Πεπερασμένου Στοιχείου με Απόσβεση για τη Στατική και Δυναμική Ανάλυση Πτερυγίων Ανεμογεννητριών" είναι η ανάπτυξη προτύπων πεπερασμένων στοιχείων με απόσβεση ικανών να προβλέπουν τη στατική και δυναμική απόκριση δοκών και πτερυγίων από σύνθετα υλικά. Η εργασία επικεντρώνεται σε δύο κύριες κατευθύνσεις, που αφορούν τόσο την εισαγωγή νέων όρων στο μητρώο απόσβεσης ενός πεπερασμένου στοιχείου δοκού, όσο και την ανάπτυξη ενός μη-γραμμικού κώδικα πεπερασμένου στοιχείου για τη μελέτη της μη-γραμμικής συμπεριφοράς δοκών και πτερυγίων από σύνθετα υλικά που υπόκεινται σε μεγάλες μετατοπίσεις και περιστροφές. Στο πρώτο μέρος της διατριβής μελετάται η επίδραση της σύζευξης, λόγω της ανισοτροπίας του σύνθετου υλικού, τόσο στη στατική απόκριση όσο και στα μορφικά χαρακτηριστικά κατασκευών από σύνθετα υλικά, διαφόρων διατομών και γεωμετριών. Διατυπώνονται νέοι όροι απόσβεσης που εκφράζουν την εν λόγω σύζευξη και οι οποίοι καθιστούν το γραμμικό πεπερασμένο στοιχείο δοκού πιο πλήρες στην επίλυση προβλημάτων όπου η σύζευξη υλικού επηρεάζει τη συμπεριφορά της κατασκευής. Στο δεύτερο και πλέον σημαντικό μέρος της παρούσας διατριβής αρχικά περιγράφεται το θεωρητικό υπόβαθρο για την πρόβλεψη της μη-γραμμικής δυναμικής απόσβεσης λεπτών δοκών κατασκευασμένα από σύνθετα υλικά οι οποίες υπόκεινται σε μεγάλα συν-επίπεδα εφελκυστικά φορτία ή φορτία λυγισμού. Αναπτύσσεται νέο πεπερασμένο στοιχείο ικανό να περιγράψει την επίδραση της γεωμετρικής μη-γραμμικότητας στην απόσβεση και τη δυσκαμψία της δοκού. Εφαλτήριο για την ανάπτυξη αυτής της μεθοδολογίας ήταν η ανάγκη της πρόβλεψης της δυναμικής απόσβεσης σε κατασκευές από σύνθετα υλικά με πιο πολύπλοκη και εύκαμπτη γεωμετρία, όπως αυτή των πτερυγίων ανεμογεννητριών. Η ανάπτυξη του μη-γραμμικού πεπερασμένου στοιχείου ξεκινά από το επίπεδο της στρώσης του υλικού, όπου διατυπώνονται οι καταστατικές εξισώσεις θεωρώντας το ιξωδοελαστικό πρότυπο του Kelvin για το υλικό της κατασκευής. Στη συνέχεια εισάγονται οι Green-Lagrange εξισώσεις συμβιβαστού οι οποίες εκφράζουν τη γεωμετρική μη-γραμμικότητα καθώς και οι εξισώσεις κίνησης. Σε επίπεδο διατομής, οι κινηματικές υποθέσεις βασίζονται στις παραδοχές της διατμητικής θεωρίας δοκού πρώτης τάξης. Η πρόβλεψη της μη-γραμμικής απόκρισης μιας κατασκευής από σύνθετα υλικά επιτυγχάνεται μέσω της προσομοίωσης της με έναν επαρκή αριθμό πεπερασμένων στοιχείων. Στο εσωτερικό κάθε στοιχείου οι παραμορφώσεις προσεγγίζονται από γραμμικές συναρτήσεις μορφής, οι οποίες οδηγούν στη μητρωική μορφή των μη-γραμμικών εξισώσεων του συστήματος. Λόγω του γεγονότος ότι οι εξισώσεις αυτές εξαρτώνται από τη λύση, δεν μπορούν να λυθούν απευθείας κάτι που καθιστά αναγκαία τη χρήση μιας σταδιακής-επαναληπτικής τεχνικής. Στην παρούσα διατριβή εισάγεται στο λύτη του μη-γραμμικού κώδικα η Newton-Raphson τεχνική. Το επόμενο βήμα αφορά τη σύνθεση των ολικών δομικών μητρών του συστήματος και την εφαρμογή των συνοριακών συνθηκών. Σε κάθε επανάληψη λαμβάνει χώρα η επίλυση των γραμμικοποιημένων εξισώσεων και ο υπολογισμός των πραγματικών και εφαπτομενικών μη-γραμμικών μητρώων δυσκαμψίας και απόσβεσης της κατασκευής, τα οποία τελικώς επιλύονται με τη μέθοδο της αριθμητικής ολοκλήρωσης κατά Gauss. Το πεπερασμένο στοιχείο δοκού εξελίχθηκε περαιτέρω ώστε να συμπεριλάβει τη μη-γραμμική ανάλυση μεγάλων λεπτότοιχων κατασκευών από σύνθετα υλικά, όπως αυτά των πτερυγίων ελικοπτέρων και ανεμογεννητριών. Η εισαγωγή της πλήρους έκφρασης της αξονικής μη-γραμμικής Green-Lagrange παραμόρφωσης στη διατύπωση των κινηματικών υποθέσεων οδηγεί στην πλήρη έκφραση των πραγματικών και εφαπτομενικών δομικών μητρών της κατασκευής. Οι νέοι μη-γραμμικοί όροι δυσκαμψίας και απόσβεσης πρώτης και δεύτερης τάξης μπορούν να περιγράψουν την επίδραση των εσωτερικών εφελκυστικών τάσεων στα μορφικά χαρακτηριστικά δοκών και πτερυγίων. Το μη-γραμμικό πεπερασμένο στοιχείο είναι ικανό να χαρακτηρίσει τη στατική συμπεριφορά και την αποσβενυμένη ταλάντωση δοκών από σύνθετα υλικά. Η επαλήθευση του μη-γραμμικού κώδικα πραγματοποιήθηκε μέσω μιας σειράς πειραματικών μετρήσεων που αφορούσαν τη μέτρηση της φυσικής συχνότητας και της μορφικής απόσβεσης σε λεπτές δοκούς από σύνθετα υλικά τόσο σε εφελκυσμό όσο και σε συνθήκες λυγισμού. Τα πειραματικά αποτελέσματα έρχονται σε πολύ καλή συμφωνία με τις θεωρητικές προβλέψεις του κώδικα κάτι που εξασφαλίζει την αξιοπιστία του μη-γραμμικού πεπερασμένου στοιχείου.

L’analyse dynamique des structures de grandes dimensions incluant de nombreux paramètres incertains et des non-linéarités localisées ou réparties peut être numériquement prohibitive. Afin de surmonter ce problème, des modèles d’approximation peuvent être développés pour reproduire avec précision et à faible coût de calcul la réponse de la structure.L’objectif de la première partie de ce mémoire est de développer des modèles numériques robustes vis-à-vis des modifications structurales (non-linéarités localisées, perturbations ou incertitudes paramétriques) et « légers » au sens de la réduction de la taille. Ces modèles sont construits, selon les approches de condensation directe et par synthèse modale, en enrichissant des bases de réduction tronquées, modale et de Craig-Bampton respectivement, avec des résidus statiques prenant compte des modifications structurales. Pour propager les incertitudes, l’accent est mis particulièrement sur la méthode du chaos polynomial généralisé. Sa combinaison avec les modèles réduits ainsi obtenus permet de créer des métamodèles mono et bi-niveaux, respectivement. Les deux métamodèles proposés sont comparés à d’autres métamodèles basés sur les méthodes du chaos polynomial généralisé et du Latin Hypercube appliquées sur des modèles complets et réduits. Les métamodèles proposés permettent d’approximer les comportements structuraux avec un coût de calcul raisonnable et sans perte significative de précision.La deuxième partie de ce mémoire est consacrée à l’analyse dynamique des structures périodiques non-linéaires en présence des imperfections : perturbations des paramètres structuraux ou incertitudes paramétriques. Deux études : déterministe ou stochastique, respectivement, sont donc menées. Pour ces deux configurations, un modèle analytique discret générique est proposé. Il consiste à appliquer la méthode des échelles multiples et la méthode de perturbation pour résoudre l’équation de mouvement et de projecter la solution obtenue sur des modes d’ondes stationnaires. Le modèle proposé conduit à un ensemble d’équations algébriques complexes couplées, fonctions du nombre et des positions des imperfections dans la structure. La propagation des incertitudes à travers le modèle ainsi construit est finalement assurée par les méthodes du Latin Hypercube et du chaos polynomial généralisé. La robustesse de la dynamique collective vis-à-vis des imperfections est étudiée à travers l’analyse statistique de la dispersion des réponses fréquentielles et des bassins d’attraction dans le domaine de multistabilité. L’étude numérique montre que la présence des imperfections dans une structure périodique renforce sa non-linéarité, élargit son domaine de multistabilité et génère une multiplicité de branches multimodale. / Dynamic analysis of large scale structures including several uncertain parameters and localized or distributed nonlinearities may be computationally unaffordable. In order to overcome this issue, approximation models can be developed to reproduce accurately the structural response at a low computational cost.The purpose of the first part of this thesis is to develop numerical models which must be robust against structural modifications (localized nonlinearities, parametric uncertainties or perturbations) and reduce the size of the initial problem. These models are created, according to the direct condensation and the component mode synthesis, by enriching truncated reduction modal bases and Craig-Bampton transformations, respectively, with static residual vectors accounting for the structural modifications. To propagate uncertainties through these first-level and second-level reduced order models, respectively, we focus particularly on the generalized polynomial chaos method. This methods combination allows creating first-level and second-level metamodels, respectively. The two proposed metamodels are compared to other metamodels based on the polynomial chaos method and Latin Hypercube method applied on reduced and full models. The proposed metamodels allow approximating the structural behavior at a low computational cost without a significant loss of accuracy.The second part of this thesis is devoted to the dynamic analysis of nonlinear periodic structures in presence of imperfections: parametric perturbations or uncertainties. Deterministic or stochastic analyses, respectively, are therefore carried out. For both configurations, a generic discrete analytical model is proposed. It consists in applying the multiple scales method and the perturbation theory to solve the equation of motion and then on projecting the resulting solution on standing wave modes. The proposed model leads to a set of coupled complex algebraic equations, depending on the number and positions of imperfections in the structure. Uncertainty propagation through the proposed model is finally done using the Latin Hypercube method and the generalized polynomial chaos expansion. The robustness the collective dynamics against imperfections is studied through statistical analysis of the frequency responses and the basins of attraction dispersions in the multistability domain. Numerical results show that the presence of imperfections in a periodic structure strengthens its nonlinearity, expands its multistability domain and generates a multiplicity of multimodal branches.

The goal of this work is to develop and demonstrate a comprehensive analysis of single and multi-body composite strip-beam systems using an asymptotically-correct geometrically nonlinear theory. The comprehensiveness refers to the two distinguishing features of this work, namely the unified framework for the analysis and the inclusion of the usually ignored cross-sectional nonlinearities in thin-beam and multi-beam analyses. The first part of this work stitches together an approach to analyse generally anisotropic composite beams. Based on geometrically exact nonlinear elasticity theory, the nonlinear 3-D beam problem splits into either a linear (conventionally considered) or nonlinear (considered in this work) 2-D analysis of the beam cross-section and a nonlinear 1-D analysis along the beam reference curve. The two sub-tasks of this work (viz. nonlinear analysis of the beam cross-section and nonlinear beam analysis) are accomplished on a single platform using an object-oriented framework. First, two established nonlinear cross-sectional analyses (numerical and analytical), both based on the Variational-Asymptotic Method (VAM), are invoked. The numerical analysis is capable of treating cross-sections of arbitrary geometry and material distributions and can capture certain nonlinear effects such as the trapeze effect. The closed-form analytical analysis is restricted to thin rectangular cross-sections for generally anisotropic composites but captures ALL cross-sectional nonlinearities, and not just the well-known Brazier and trapeze effects. Second, the well-established geometrically-exact nonlinear 1-D governing equations along the beam reference curve, after being generalized to utilize the expressions for nonlinear stiffness matrix, are solved using the mixed variational finite element method. Finally, local 3-D stress, strain and displacement fields for representative sections in the beam are recovered, based on the stress resultants from the 1-D global beam analysis. This part of the work is then validated by applying it to an initially twisted cantilevered laminated composite strip under axial force. The second part is concerned with the dynamic analysis of nonlinear multi-body systems involving elastic strip-like beams made of laminated, anisotropic composite materials using an object-oriented framework. In this work, unconditionally stable time-integration schemes presenting high-frequency numerical dissipation are used to solve the ensuing governing equations. The codes developed based on such time-integration schemes are first validated with the literature for two standard test cases: non-linear spring mass oscillator and pendulum. In order to apply the comprehensive analysis code thus developed to a multi-body system, the four-bar mechanism is chosen as an example. All component bars of the mechanism have thin rectangular cross-sections and are made of fiber reinforced laminates of various types of layups. They could, in general, be pre-twisted and/or possess initial curvature, either by design or by defect. They are linked to each other by means of revolute joints. Each component of the mechanism is modeled as a beam based on the first part of this work. Results from this analysis are compared with those available in the literature, both theoretical and experimental. The margins between the linear and non-linear results are evaluated specifically due to the cross-sectional nonlinearities and shown to vary with stacking sequences. This work thus demonstrates the importance of geometrically nonlinear cross-sectional analysis of certain composite beam-based four-bar mechanisms in predicting system dynamic characteristics. To enable graphical visualization, the behavior of the four-bar mechanism is also observed by using commercial software (I-DEAS + NASTRAN + ADAMS). Finally, the component-laminate load-carrying capacity is estimated using the Tsai-Wu-Hahn failure criterion for various layups and the same criterion is used to predict the first-ply-failure and the mechanism as a whole.

In vielen industriellen Anwendungen sind Dichtungen im Einsatz. Vergleicht man den Preis mit dem eines Gesamtsystems, in denen Dichtungen verwendet werden, so sind Dichtungen verhältnismäßig günstig. Jedoch führt ein Versagen von Dichtungen meist zu schwerwiegenden Konsequenzen. Dichtungen sind komplexe Subsysteme und ihre Auslegung erfordert umfangreiche Kenntnisse im Bereich Materialmodellierung, Belastung und Versagenskriterien. Die heutige Simulationstechnologie ermöglicht einen parametrischen Workflow für die Berechnung des Verhaltens von Dichtungen mit den auftretenden Effekten wie nichtlinearem Materialverhalten, wechselnden Kontaktbedingungen und Flüssigkeitsunterwanderung bei Druck. Als ein führendes Simulationswerkzeug für diese physikalische Fragestellung wird ANSYS Mechanical für die Auslegung herangezogen. Desweiteren kann das Verständnis für das Produkt erhöht werden, was zu einer Verbesserung der Funktionalität und der Zuverlässigkeit führt. Versuchsdaten können als Spannungs-Dehnungskurven in ANSYS importiert werden, welche das Materialverhalten des hyperelastischen Werkstoffs mit traditionellen Materialmodellen wie Mooney Rivlin, Ogden and Yeoh oder einer neueren Formulierung, der Antwortfunktionsmethode, widerspiegeln. Robuste Kontakttechnologien beschleunigen die Simulation und Entwicklungszeit-Berechnungszeiten und gewährleisten ein genaues Verhalten des Simulationsmodells. Insbesondere bei Dichtungen ist die druckbeaufschlagte Fläche in 2D und 3D Anwendungen von Bedeutung. ANSYS berechnet diese automatisch in Abhängigkeit des aktuellen Kontaktzustandes. Diese benutzerfreundliche Unterstützung führt zu einer höheren Genauigkeit des Simulationsergebnisses, da ein manuelles Schätzen der Druckflächen entfällt. Mit einem parametrischen und durchgängigen Ansatz innerhalb von ANSYS Workbench, beginnend bei der CAD-Geometrie, über die Vernetzung, Material- und Randbedingungsdefinition und Lösung. können eine Reihe von Varianten in kurzer Zeit berechnet werden. Neben einem besseren Verständnis für das Produkt hilft dies dem Ingenieur Änderungen vorzunehmen, was zu exakten und aussagekräftigen Ergebnissen führt. Desweiteren kann der Einfluss von Unsicherheiten berücksichtigt werden, sodass der Berechnungsingenieur fernab von idealen Bedingungen robuste und zuverlässige Dichtungen entwickeln kann.