Non-ideal magnetohydrodynamic (MHD) effects may play a significant role in determining the dynamics, thermal properties, and observational signatures of radiatively inefficient accretion flows onto black holes. In particular, particle acceleration during magnetic reconnection events may influence black hole spectra and flaring properties. We use representative general relativistic magnetohydrodynamic (GRMHD) simulations of black hole accretion flows to identify and explore the structures and properties of current sheets as potential sites of magnetic reconnection. In the case of standard and normal evolution (SANE) disks, we find that in the reconnection sites, the plasma beta ranges from 0.1 to 1000, the magnetization ranges from 10(-4) to 1, and the guide fields are weak compared with the reconnecting fields. In magnetically arrested (MAD) disks, we find typical values for plasma beta from 10(-2) to 10(3), magnetizations from 10(-3) to 10, and typically stronger guide fields, with strengths comparable to or greater than the reconnecting fields. These are critical parameters that govern the electron energy distribution resulting from magnetic reconnection and can be used in the context of plasma simulations to provide microphysics inputs to global simulations. We also find that ample magnetic energy is available in the reconnection regions to power the fluence of bright X-ray flares observed from the black hole in the center of the Milky Way.

The evolution of protoplanetary disks is believed to be driven largely by angular momentum transport resulting from magnetized disk winds and turbulent viscosity. The ionization of the disk that is essential for these processes has been thought to be due to host star coronal X-rays but could also arise from energetic particles produced by coronal flares, or traveling shock waves, and advected by the stellar wind. We have performed test-particle numerical simulations of energetic protons propagating into a realistic T. Tauri stellar wind, including a superposed small-scale magnetostatic turbulence. The isotropic (Kolmogorov power spectrum) turbulent component is synthesized along the individual particle trajectories. We have investigated the energy range [0.1-10] GeV, consistent with expectations from Chandra X-ray observations of large flares on T. Tauri stars and recent indications by the Herschel Space Observatory of a significant contribution of energetic particles to the disk ionization of young stars. In contrast with a previous theoretical study finding a dominance of energetic particles over X-rays in the ionization throughout the disk, we find that the disk ionization is likely dominated by X-rays over much of its area, except within narrow regions where particles are channeled onto the disk by the strongly tangled and turbulent magnetic field. The radial thickness of such regions is 5 stellar radii close to the star and broadens with increasing radial distance. This likely continues out to large distances from the star (10 au or greater), where particles can be copiously advected and diffused by the turbulent wind.

La prise en compte des couplages dynamiques entre les différents organes constituant une turbomachine s’inscrit dans le processus d’optimisation des designs moteur. L’amélioration des performances des turboréacteurs passe souvent par l’utilisation d’architectures multi-rotors. Dans le contexte des moteurs avec une architecture bi-rotor, des résultats d’essais expérimentaux montrent qu’il est nécessaire de considérer, dès la conception, l’influence de la dynamique de l’arbre Haute Pression (HP) sur les aubages de l’arbre Basse Pression (BP). Dans ce cadre d’étude, un premier modèle simplifié de bi-rotor aubagé est développé dans le repère tournant lié au rotor BP. Ce modèle est composé de deux rotors modélisés par des équivalents poutres - masses - ressorts et d’une roue aubagée constituée d’aubes souples modélisées par des poutres encastrées sur un disque rigide. Desnon-linéarités de type jeu radial avec contact au niveau des paliers sont également considérées et la réponse des aubes soumises à des excitations multi-fréquentielles de type balourd BP et HP est analysée. La présence de non-linéarités dans le système conduit à mettre en oeuvre des algorithmes adaptés, basés sur des techniques de résolution dans le domaine fréquentiel avec l’évaluation des efforts non-linéaires dans le domaine temporel. Afin d’avoir une meilleure description de la dynamique de la roue aubagée, une méthode spécifique de couplage est proposée, permettant de coupler un modèle réduit de roue aubagée 3D à un modèle simplifié de bi-rotor. Une démarche adaptée à la modélisation de la roue aubagée en symétrie cyclique est implémentée afin de considérer des non-linéarités de type contact en tête d’aube. La méthode de couplage proposée est ensuite illustrée sur un exemple simple puis validée dans un cadre linéaire et non-linéaire. Enfin, cette méthode de couplage est appliquée au cas d’une structure industrielle, constituée d’un modèle d’ensemble simplifié représentatif d’un moteur et d’un modèle éléments finis d’une roue de turbine BP. Les résultats obtenus mettent en évidence le couplage entre la dynamique d’ensemble et la dynamique de la roue aubagée et permettent de prédire la réponse non-linéaire des aubes de turbine BP en présence d’une excitation multi-fréquentielle, dans des configurations de co-rotation et de contra-rotation. / The design and optimization process of high efficiency turbomachinery has become a major challenge and a topical issue at both industrial and research levels. Performance improvement has motivated the use of multi-shaft architecture in engines. In the context of dual-shaft aircraft engines, the interaction between dynamics occurring within shafts and bladed disks seems to play an important role at the design stage. The present research work deals with the coupling of these components involving several unbalances in the dynamic response of blades. Within this framework, a simplified analytical model of a bladed dual-shaft developed in the rotating frame is presented. The dual-shaft is modelled by spring - mass- beam systems and connected to a bladed disk composed of a set of flexible blades modelled by Euler-Bernoulli beams clamped in a rigid disk. Nonlinearities coming from bearings are also considered and modelled as a radial clearance and contact stiffness. Considering nonlinearities requires the implementation of dedicated algorithms and specific resolution techniques in the frequency domain as well as the computation of nonlinear forces in the time domain. The nonlinear response of blades subjected to unbalances excitations is investigated and analysed. To have a finer description of the bladed disk dynamics, a specific coupling method is proposed allowing to connect a bladed disk finite element model with the simplified dual-shaft model. A cyclic symmetry approach well-suited to the nonlinear dynamics of bladed disks is developed in order to consider blade tip contact nonlinearities. Performances of the proposed method are illustrated through an academic example and validated in both linear and nonlinear settings. Eventually, the coupling technique is applied to a complex industrial case involving a classical simplified dual-shaft model and a finite element model of the low pressure turbine bladed disk. Numerical results clearly demonstrate the coupling between dynamics and enable to predict the nonlinear response of low pressure turbine blades to several unbalances, for both co-rotating and counter-rotating engines.

L’objectif de ces travaux est de prédire la stabilité dynamique d’un moteur d’avion de type turbofan, lors de contacts légers entre l’extrémité des aubes et le carter en regard. Une voix d’amélioration des performances des turboréacteurs d’avion consiste, en effet, à réduire les jeux entre les extrémités des aubes en rotation et le carter. Réduction qui s’accompagne d’une augmentation des possibilités de contact entre ces parties tournantes et fixes pouvant conduire à des phénomènes vibratoires instables. La démarche proposée s’appuie sur un modèle hybride qui introduit de manière simplifiée une roue aubagée et un carter flexible à un modèle de ligne d’arbres. Une formulation dite 3D du contact a également été implémentée, elle intègre la cinématique des différents organes du modèle et introduit de manière précise la géométrie locale de la zone de contact. Le comportement du modèle lors de contacts aubes-carter est alors étudié en distinguant les analyses menées en supposant un contact permanent et celles autorisant l’intermittence des contacts. Les résultats obtenus mettent en évidence l’importance des couplages introduits dans l’apparition des phénomènes instables et la pertinence de la formulation 3D dans la prédiction de la stabilité du système. / This work aims to predict the dynamic stability of a turbofan engine submitted to light contacts between blade tips and casing. Reducing the clearance between the rotating blades and the casing improves indeed the performances of turbomachines ; however, it also increases the possible contacts between rotating and stationary parts, which can cause unstable dynamic behavior. The approach is based on a hybrid model which introduces a simplified bladed wheel and a flexible casing to a rotor-shaft model. A 3D contact formulation has also been implemented ; it considers the model kinematic and introduces the local geometry of the contact area. The model behavior with blade-to-casing contacts is analyzed through two approaches : the first one assumes permanent contacts while the other one allows contact intermittence. The results highlight the importance of couplings in the outbreak of unstable phenomena and the relevance of the 3D contact formulation in predicting the stability of the system.

Dynamical mass estimates down to the planet-mass regime can help to understand planet formation. We present Atacama Large Millimeter/submillimeter Array (ALMA) 1.3 mm observations of FW Tau C, a proposed similar to 10M(Jup) planet-mass companion at similar to 330 au from the host binary FW Tau AB. We spatially and spectrally resolve the accretion disk of FWTau C in (CO)-C-12 (2-1). By modeling the Keplerian rotation of gas, we derive a dynamical mass of similar to 0.1 M-circle dot. Therefore, FW Tau C is unlikely a planet, but rather a low-mass star with a highly inclined disk. This also suggests that FW Tau is a triple system consisting of three similar to 0.1. M-circle dot stars.

We present multi-wavelength radio observations obtained with the VLA of the protoplanetary disk surrounding the young brown dwarf 2MASS J04442713+2512164 (2M0444) in the Taurus star-forming region. 2M0444 is the brightest known brown dwarf disk at millimeter wavelengths, making this an ideal target to probe radio emission from a young brown dwarf. Thermal emission from dust in the disk is detected at 6.8 and 9.1 mm, whereas the 1.36 cm measured flux is dominated by ionized gas emission. We combine these data with previous observations at shorter sub-mm and mm wavelengths to test the predictions of dust evolution models in gas-rich disks after adapting their parameters to the case of 2M0444. These models show that the radial drift mechanism affecting solids in a gaseous environment has to be either completely made inefficient, or significantly slowed down by very strong gas pressure bumps in order to explain the presence of mm/cm-sized grains in the outer regions of the 2M0444 disk. We also discuss the possible mechanisms for the origin of the ionized gas emission detected at 1.36 cm. The inferred radio luminosity for this emission is in line with the relation between radio and bolometric luminosity valid for for more massive and luminous young stellar objects, and extrapolated down to the very low luminosity of the 2M0444 brown dwarf.

In this thesis we explored the use of nuclear stellar discs as tracers of the merging history of early-type galaxies. These small structures, just a few tens to a few hundreds of parsecs across, are a common but poorly studied feature of early-type galaxies. They are formed during or shortly after merging events due to the infall of gas, which settles in a disc and leads to the formation of new stars. Initial simulations showed that they should not survive a following major merger and could, therefore, be used to trace the epoch when their host galaxies experienced their last major merger event. We produced the first census of nuclear discs and established that their incidence is 20%, fairly independent of the host-galaxy mass or galactic environment. Furthermore, we have more than doubled the sample of nuclear discs with known photometric properties, finding that they give a hint of possessing different characteristics from those of large, galactic discs. Using these nuclear discs as clocks for the assembly history of galaxies requires dating their stellar populations. By combining the use of integral-field spectroscopy with the a priori knowledge of the relative bulge- and disc-light contribution to the observed spectra, as determined by a photometric disc-bulge decomposition, we have shown that it is possible to reduce the degeneracies that affect the study of two superimposed populations and thus that the age of stellar discs can be measured more precisely. To illustrate our method, we present VLT-VIMOS data for NGC 4458, a low-mass slowly rotating early-type galaxy with a disc that we found to be at least 5-6 Gyr old. The presence of such an old central disc in such a small, slowly-rotating and, mostly likely, round galaxy is particularly puzzling and presents a challenge to existing models. Disc fragility is central to our studies and we have expanded the limited initial simulations to study it in more detail. By means of N-body simulations, we have reproduced the final stages of a galaxy encounter by exposing a nuclear disc rotating in the gravitational potential of its host bulge and central supermassive black hole to the impact of a secondary massive black hole. We explored not only major mergers (1:1 mass ratio), but also large minor mergers (1:5 and 1:10), across a variety of collision angles, and assessed the survival of the disc, as perceived by current observational limits, both for photometry and spectroscopy. As expected, the discs do not survive a major merger whereas it is in general possible to detect their presence after a 1:5 or 1:10 encounter, in particular when looking at kinematic signatures with spectroscopy. This thesis has demonstrated that nuclear discs constitute both a common and accurate tool for constraining the assembling history of nearby early-type galaxies. The advent of more sensitive integral-field spectrographs, such as MUSE, will make measuring the stellar age of nuclear discs not only more precise, but also more economical in terms of telescope time. This will allow embarking on a more systematic age dating campaign for nuclear stellar discs across a wider range of type, mass and galactic environments for their host galaxies. Combining such a census with a larger set of numerical simulations aimed at calibrating better the range of merger event that would erase any photometric or kinematic signature of a nuclear disc, should finally allow us to put firm constraints on the merging history of early-type galaxies.

Massive circumstellar disks are prone to gravitational instabilities, which trigger the formation of spiral arms that can fragment into bound clumps under the right conditions. Two-dimensional simulations of self-gravitating disks are useful starting points for studying fragmentation because they allow high-resolution simulations of thin disks. However, convergence issues can arise in 2D from various sources. One of these sources is the 2D approximation of self-gravity, which exaggerates the effect of self-gravity on small scales when the potential is not smoothed to account for the assumed vertical extent of the disk. This effect is enhanced by increased resolution, resulting in fragmentation at longer cooling timescales beta. If true, it suggests that the 3D simulations of disk fragmentation may not have the same convergence problem and could be used to examine the nature of fragmentation without smoothing self-gravity on scales similar to the disk scale height. To that end, we have carried out local 3D self-gravitating disk simulations with simple beta cooling with fixed background irradiation to determine if 3D is necessary to properly describe disk fragmentation. Above a resolution of similar to 40 grid cells per scale height, we find that our simulations converge with respect to the cooling timescale. This result converges in agreement with analytic expectations which place a fragmentation boundary at beta(crit) = 3.

Accretion discs efficiently transport angular momentum by a wide variety of as yet imperfectly understood mechanisms, with profound implications for the disc lifetime and planet formation. We discuss two different methods of angular momentum transport: first, generation of acoustic waves by mixing of inertial waves, and second, the generation of a self-sustaining magnetic field via the magnetorotational instability (MRI) which would be a source of dissipative turbulence. Previous local simulations of the MRI have shown that the dynamo changes character on addition of vertical stratification. We investigate numerically 3D hydrodynamic shearing waves with a conserved Hermitian form in an isothermal disc with vertical gravity, and describe the associated symplectic structure. We continue with a numerical investigation into the linear evolution of the MRI and the undular magnetic buoyancy instability in isolated flux regions and characterise the resultant quasi-linear EMFs as a function of height above the midplane. We combine this with an analytic description of the linear modes under an assumption of a poloidal-toroidal scale separation. Finally, we use RAMSES to perform full MHD simulations in a zero net flux shearing box, followed by spatial and a novel temporal averaging to reveal the essential structure of the dynamo. We find that inertial modes may be efficiently converted into acoustic modes for "bending waves", despite a fundamental ambiguity in the inertial mode structure. With our linear MRI and the undular magnetic buoyancy modes we find the localisation of the instability high in the atmosphere becomes determined by magnetic buoyancy rather than field strength for small enough azimuthal wavenumber, and that the critical Alfven speed below which the dynamo can operate increases with increasing distance from the midplane. We calculate analytically quasi-linear EMFs which predict both a vertical propagation of toroidal field and a method for creation of radial field. From our fully nonlinear calculations we find an electromotive force in phase with the toroidal field, which is itself 3π/2 out of phase with the radial (sheared) field at the midplane, and good agreement with our quasi-linear analytics. We have identified an efficient mechanism for generating acoustic waves in a disc. In our investigation of the accretion disc dynamo, we have reproduced analytically the EMFs calculated in our simulations, given arguments based on the phase of relevant quantities, several correlation integrals and the scalings suggested by our analytic work. Our analysis contributes significantly to an explanation for the dynamo in an accretion disc.

Estrelas Be são reconhecidas pela sua rápida rotação e pulsação não radial. São as únicas estrelas da Sequência Principal que apresentam discos circunstelares, os quais são formados por meio de processos ainda não completamente compreendidos. A modelagem das forças que atuam neste sistema conduz a previsões teóricas sobre a estrutura do disco que podem ser comparadas com dados observacionais. Podemos estudar as propriedades físicas dos discos de estrelas Be pelo efeito que a luz estelar sofre ao passar por eles, por exemplo, modelando a transferência radiativa. Neste ponto, a polarização surge como uma ferramenta muito útil para a investigação destes discos, permitindo a determinação de quantidades físicas importantes do sistema, como a densidade numérica de partículas e o ângulo de inclinação. Uma variabilidade intrigante observada em estrelas Be é a transição aperiódica entre uma fase B normal (sem disco) e uma fase Be (com disco). Estudos de monitoramento recentes encontraram, a partir da análise da polarização intrínseca decorrente da transição entre estas fases, uma relação significante entre a mudança da polarização através do salto de Balmer versus a polarização na banda V, fazendo surgir uma estrutura em loop como função do tempo, no assim denominado Diagrama Cor-Polarização. Neste trabalho, apresentamos uma análise do Diagrama Cor-Polarização por meio de modelos diversos. Fazemos uso do Disco de Decréscimo Viscoso que é o paradigma atual para explicar a formação e evolução dos discos de estrelas Be. Com isso, visamos determinar como a polarimetria pode contribuir para a compreensão dos mecanismos fundamentais envolvidos no processo de formação e dissipação do disco. / Be stars are recognized by their rapid rotation and non-radial pulsation. They are the only stars in the Main Sequence that have circumstellar disks that are formed by processes not yet fully understood. The modeling of the forces acting on this system leads to theoretical predictions about the structure of the disk that can be compared to observational data. We can study physical the properties of Be disks by modeling how stellar light is reprocessed by them. This requires solving the detailed radiative transfer problem involved. In this point, the study of polarization arise as a useful tool to investigate these disks, allowing for the determination of important physical quantities of the system, such as the particle number density and inclination angle. An intriguing variability observed in Be stars is the aperiodic transition between a B normal phase (without disk) to a Be phase (with disk). Recent monitoring studies found, from the analysis of the intrinsic polarization arising of the transition between these phases, a significant relation between the polarization change through the Balmer jump versus the V-Band polarization, giving rise to a loop structure as a function of time, in the so-called Color-Polarization Diagram. This work presents an analysis of the Color-Polarization Diagram by several models. We make use of the Viscous Decretion Disk Model, which assumes the existence of some injection mechanism of material at keplerian velocities in the disk base, where the turbulent viscosity acts carrying angular momentum from de inner parts to the outer regions. With this, we aimed to extend our knowledge about the fundamental mechanisms involved in the formation and dissipation processes of the disk.