Práce se zabývá vlivem disipativních proces· v mezních vrstvách, konkrétně vlivem viskozity tekutiny a dále vlivem difúze vodní páry ve vzduchu na vznik sil p·sobících na obtékaná tělesa. Z Croccovy věty plyne, že gradient entropie má sv·j podíl na vzniku cirkulace a d·sledkem toho i na vznik vztlakové síly. Práce zkoumá vznik vztlakové síly při obtékání rotujícího válce, v soft- waru FEniCS modeluje tuto situaci, porovnává s modelem nevazké nestlačitelné tekutiny a následně i s experimentem provedeným Ing. Zdeňkem Trávníčkem a Ing. Zuzanou Broučkovou. Taktéž zkoumá vznik vztlakové síly při vypařování na horní straně válce a srovnává s případem rotujícího válce. 1

Magnetic fluid flow by thermomagnetic convection with and without buoyancy was studied in experiments and computational simulations. A mineral oil based ferro magnetic fluid was subjected to varying magnetic fields to induce thermomagnetic convection. As such fluids are mainly developed to increase heat transfer for cooling the fundamental effects on magnetic fluid flow was investigated using various magnetic field distributions. Computational simulations of natural and thermomagnetic convection are based on a Finite-Element technique and considered a constant magnetic field gradient, a realistic magnetic field generated by a permanent magnet and alternating magnetic fields. The magnetic field within the fluid domain was calculated by the magneto-static Maxwell equations and considered in an additional magnetic body force known as the Kelvin body force by numerical simulations. The computational model coupled the solutions of the magnetic field equations with the heat and fluid flow equations. Experiments to investigate thermomagnetic convection in the presence of terrestrial gravity used infrared thermography to record temperature fields that are validated by a corresponding numerical analysis. All configurations were chosen to investigate the response of the magnetic fluid to the applied body forces and their competition by varying the magnetic field intensity and its spatial distribution. As both body forces are temperature dependent, situations were analysed numerically and experimentally to give an indication of the degree by which heat transfer may be enhanced or reduced. Results demonstrate that the Kelvin body force can be much stronger than buoyancy and can induce convection where buoyancy is not able to. This was evident in a transition area if parts of a fluid domain are not fully magnetically saturated. Results for the transition from natural convection to thermomagnetic convection suggest that the domain of influence of the Kelvin body force is aligned with the dominance of the respective body force. To characterise the transition a body force ratio of the Kelvin body force to buoyancy was developed that identified the respective driving forces of the convection cells. The effects on heat transfer was quantified by the Nusselt number and a suitable Rayleigh number. A modified Rayleigh number was used when both body forces were active to define an effective body force by taking the relative orientation of both forces into account. Results for the alternating magnetic field presented flow fields that altered with the frequency of the applied magnetic field but with varying amplitude. This affected the heat transfer that alternated with the frequency but failed to respond instantaneously and a phase lag was observed which was characterised by three different time scales.

<p>High-precision differential cross section data of the neutron-deuteron elastic scattering reaction at 95 MeV are presented. The neutron-proton scattering differential cross section was also measured and used as a reference to allow an accurate absolute normalization of the neutron-deuteron data.</p><p>Two multi-detector arrays were used, MEDLEY and SCANDAL, at the neutron beam facility at The Svedberg Laboratory in Uppsala. Three different configurations of the detectors allowed to perform three independent measurements. The first experiment involved detecting recoil deuterons from thin deuterated polyethylene targets with the MEDLEY setup and allowed a large angular coverage. In the second experiment, high-precision data were obtained at neutron backward angles, using the SCANDAL setup with the same technique. For the third experiment, data were obtained in the forward angular range using the SCANDAL setup with a technique where neutrons scattered on heavy water were detected by neutron-proton conversion in plastic scintillators and tracking the protons through the detectors. Events from elastic neutron-deuteron scattering were identified in the data analysis, and differential cross sections were obtained after applying corrections and evaluating systematic uncertainties due to effects which could affect the shape or the absolute normalization of the data.</p><p>The results are compared with modern Faddeev calculations using realistic nucleon-nucleon potentials combined with three-nucleon interactions. The effects of three-nucleon forces are expected to increase the differential cross section by about 30% in the region of the minimum. The data agree with this prediction, thus providing evidence for three-nucleon force effects.</p>

High-precision differential cross section data of the neutron-deuteron elastic scattering reaction at 95 MeV are presented. The neutron-proton scattering differential cross section was also measured and used as a reference to allow an accurate absolute normalization of the neutron-deuteron data. Two multi-detector arrays were used, MEDLEY and SCANDAL, at the neutron beam facility at The Svedberg Laboratory in Uppsala. Three different configurations of the detectors allowed to perform three independent measurements. The first experiment involved detecting recoil deuterons from thin deuterated polyethylene targets with the MEDLEY setup and allowed a large angular coverage. In the second experiment, high-precision data were obtained at neutron backward angles, using the SCANDAL setup with the same technique. For the third experiment, data were obtained in the forward angular range using the SCANDAL setup with a technique where neutrons scattered on heavy water were detected by neutron-proton conversion in plastic scintillators and tracking the protons through the detectors. Events from elastic neutron-deuteron scattering were identified in the data analysis, and differential cross sections were obtained after applying corrections and evaluating systematic uncertainties due to effects which could affect the shape or the absolute normalization of the data. The results are compared with modern Faddeev calculations using realistic nucleon-nucleon potentials combined with three-nucleon interactions. The effects of three-nucleon forces are expected to increase the differential cross section by about 30% in the region of the minimum. The data agree with this prediction, thus providing evidence for three-nucleon force effects.

No description available.

No description available.

No description available.

Cette thèse étudie des méthodes de conception aérodynamique pour les avions de ligne de demain. A l'heure actuelle, les avions de ligne sont en général conçus de manière à ce que les moteurs, conçus séparément du reste de l'aéronef, n'interagissent que très peu avec la cellule de l'avion (la voilure, le fuselage,...). Pour diminuer la consommation de carburant, de nouveaux concepts comme l'ingestion de couche limite émergent, dans lesquels l'avion est conçu pour tirer profit des interactions aérodynamiques qui peuvent s'établir entre le moteur et la cellule de l'avion sur certaines configurations. Il devient alors nécessaire de simuler ces interactions pour s'assurer que le bénéfice pour l'avion en termes de consommation de carburant est réel. La méthode développée dans cette thèse a pour objectif de rendre possible la simulation de ces interactions, à un coût de calcul qui reste acceptable. La soufflante, qui est l'élément du moteur le plus à même d'interagir avec l'avion, est modélisée à l'aide d'un champ de force qui reproduit son aspiration de l'écoulement d'air. Cette approche permet de reproduire les interactions aérodynamiques entre l'avion et ses moteurs à un coût 50 fois inférieur à celui d'une simulation complète du moteur et de l'avion, ce qui permettra en pratique d'optimiser les lignes aérodynamiques des futurs avions. / This work explores new méthodologies for the aerodynamic conception of future commercial aircraft. In general, commercial aircraft are designed to limit aerodynamic interactions between the engines and the airframe. New aircraft concepts such as boundary layer ingestion are now studied, in which the aircraft is designed to take advantage of these interactions. It is then necessary to be able to simulate these interactions to ensure that real benefits in terms of aircraft fuel burn are possible. The methodology that is developed in the présent thesis aims at enabling the simulation of these aerodynamic interactions at affordable computational cost. The fan, which is the part of the engine the most likely to internet with the airframe, is modeled using a force field that reproduces the suction of the air inside the engine. This approach allows to reproduce fan- airframe interactions at a fraction of the cost of a complété simulation of the aircraft and the engines, and enable the practical optimization of the aerodynamic performance of future aircraft.

O presente trabalho apresenta um método de interação fluido estrutura chamado de fronteira virtual para modelagem de escoamentos sobre geometrias complexas. A principal característica do método é que a condição de contorno interna da superfície do corpo imerso é obtida usando uma função que calcula a força de campo das equações de momentum. Este método de cálculo da força de campo é conhecido como método direto. A principal vantagem do método da fronteira virtual é que se trabalha com duas malhas, uma cartesiana (também conhecida como euleriana) para o fluido e outra curvilínea (também conhecida como malha lagrangiana) para o objeto imerso. Como estas malhas são independentes uma da outra, os algoritmos de geração das malhas são simples em comparação com outros métodos de geração de malha, nos quais a esta precisa se adaptar ao corpo que se quer analisar. Nos métodos de body fitted, algoritmos complexos com custo computacional muito elevado são necessários para se trabalhar com geometrias complexas. Neste trabalho, o método de fronteira virtual é desenvolvido para trabalhar escoamentos compressíveis não viscosos em duas dimensões sobre geometrias complexas, e testado sobre um cilindro de seção circular e sobre uma série de aerofólios NACA da série 4. O regime de principal interesse é o regime transônico, em particular para velocidades abaixo da velocidade do som (Mach entre 0,8 e 1). No entanto, resultados para outros regimes de escoamento (subsônico incompressível, subsônico compressível, transônico e supersônico) também são mostrados. / This work presents a fluid-structure interaction method, known as virtual boundary method, developed for flow modeling, over complex geometries. The main characteristic of the method is that, the internal boundary condition at the surface are created using a function, that compute the body force from the momentum equations. This method is known as direct method. The virtual boundary method main advantage is the easy and quick mesh generation. Due the use two different grids, a Cartesian grid for the fluid and a curvilinear grid for the body that interact with the fluid. As these two grids are independent one of other, the grid generation algorithms are easier compared with other methods on CFD as the body fitted. Where the grid has to be adapted to the body of interest. The body fitted methods are implemented with complex algorithms, and high computational cost, especially when complex geometries are analyzed. In this work, the virtual boundary method are developed in order to work with in viscid compressible flow in two dimensions over complex geometries, tested over a cylinder and several NACA series 4 and critical airfoils. The regime of interest is the transonic, below the speed of sound, nevertheless results for several flow regimens (incompressible subsonic, compressible subsonic, transonic and supersonic) are also presented.

O presente trabalho apresenta um método de interação fluido estrutura chamado de fronteira virtual para modelagem de escoamentos sobre geometrias complexas. A principal característica do método é que a condição de contorno interna da superfície do corpo imerso é obtida usando uma função que calcula a força de campo das equações de momentum. Este método de cálculo da força de campo é conhecido como método direto. A principal vantagem do método da fronteira virtual é que se trabalha com duas malhas, uma cartesiana (também conhecida como euleriana) para o fluido e outra curvilínea (também conhecida como malha lagrangiana) para o objeto imerso. Como estas malhas são independentes uma da outra, os algoritmos de geração das malhas são simples em comparação com outros métodos de geração de malha, nos quais a esta precisa se adaptar ao corpo que se quer analisar. Nos métodos de body fitted, algoritmos complexos com custo computacional muito elevado são necessários para se trabalhar com geometrias complexas. Neste trabalho, o método de fronteira virtual é desenvolvido para trabalhar escoamentos compressíveis não viscosos em duas dimensões sobre geometrias complexas, e testado sobre um cilindro de seção circular e sobre uma série de aerofólios NACA da série 4. O regime de principal interesse é o regime transônico, em particular para velocidades abaixo da velocidade do som (Mach entre 0,8 e 1). No entanto, resultados para outros regimes de escoamento (subsônico incompressível, subsônico compressível, transônico e supersônico) também são mostrados. / This work presents a fluid-structure interaction method, known as virtual boundary method, developed for flow modeling, over complex geometries. The main characteristic of the method is that, the internal boundary condition at the surface are created using a function, that compute the body force from the momentum equations. This method is known as direct method. The virtual boundary method main advantage is the easy and quick mesh generation. Due the use two different grids, a Cartesian grid for the fluid and a curvilinear grid for the body that interact with the fluid. As these two grids are independent one of other, the grid generation algorithms are easier compared with other methods on CFD as the body fitted. Where the grid has to be adapted to the body of interest. The body fitted methods are implemented with complex algorithms, and high computational cost, especially when complex geometries are analyzed. In this work, the virtual boundary method are developed in order to work with in viscid compressible flow in two dimensions over complex geometries, tested over a cylinder and several NACA series 4 and critical airfoils. The regime of interest is the transonic, below the speed of sound, nevertheless results for several flow regimens (incompressible subsonic, compressible subsonic, transonic and supersonic) are also presented.