Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

The transition of boundary layer from laminar to turbulent is more probable to happen across various components of a turbomachine. If narrowed down to a case of low-pressure turbine cascade, the transition of flow in boundary layer is due to separation induction. Due to this, blade losses are observed and it depends on various physical characteristics such as size and the length of the separation bubble. The physical characteristics can vary based on the variation of flow Reynolds number, expansion ration of the flow and the inlet flow turbulence intensity. Such characteristics can be studied using computational techniques, CFD analysis. Various computational techniques can be used for this scope of study, ranging from Reynolds Averaged Navier Stokes to Direct Numerical Simulation modelling, and each technique have its own set of advantages and disadvantages. 

The objective of this thesis is to present and acknowledge on how these separation bubble physical characteristics vary due to the variation of flow expansion ratio and variation of exit isentropic Reynolds number for a SPLEEN blade cascade. Direct Numerical Simulation technique will be adapted for this thesis, due to its ability to accurately predict separation bubble location and size but compromising to relatively high computational power requirement.

Low pressure turbine --- Flow separation bubble --- Boundary layer --- Flow transition --- Turbulence --- Reynolds stress --- Direct Numerical Simulation --- Discontinous Galerkin method --- Ingénierie, informatique & technologie > Ingénierie aérospatiale

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Lately, the interest towards drones for both commercial and civilian application has significantly
grown. Most of the current solutions are of the family of multicopters. Their intrinsic nature,
however, makes them not suitable for long range missions, where fixed wing configurations are
usually preferred. This thesis work consists on the conceptual design of an autonomous fixed
wing UAV, able of carrying and delivering a configurable payload.
In this work, the method proposed by Roskam, Torenbeek and Gundlach is followed throughout
the design process. Initial and conceptual design phases are carried out. First, the main mass
and power properties of the aircraft have been estimated making use of linear regressions
based on the available data. After a configuration study, more advanced methods, presented
in the literature have been employed to obtain more accurate estimations of the required parameters.
The Blended Wing Body configuration was proven to be the most aerodynamically
efficient and thus it was object of an optimisation process.
Finally the control surfaces have been designed, the Penaud diagrams presented, The material
selection process carried out making use of the Enhanced Digital Logic method; a preliminary
design of the cargo bay is proposed, as well as a general overview of the take off and landing
systems

Ingénierie, informatique & technologie > Ingénierie aérospatiale

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

This master thesis was led in KTH, Kungliga Tekniska Högskolan, in the Maritime Robotics Laboratory. The topic of the master thesis is the performance analysis of a reduced wind powered cargo model. This boat is a reduced version of the full-scale concept called OceanBird, part of the Wind Powered Car Carrier (wPCC) project. This project is a collaboration between the shipping company Wallenius Marine, the marine technology company SSPA and the Royal Institute of Technology KTH, which aims to cut 90 % of the fuel emissions of such cargo. The full-size project is a car-carrier that will be 200 m long and have five 80 m high rigid sails able to carry 7000 cars on board. KTH has the privilege of having a reduced model, exactly 1/30 of the full scale Oceanbird, to test and study the boat. This model is 7 meters long and features four wings of 4 meters high made out of fiberglass that propel the boat by creating lift. Such research is very new and thus does not provide a lot of comparison with previous studies. The performance of the boat is evaluated by its ability to manoeuvre in varying weather conditions. Different criteria are analysed in order to help assess the performance of the full-scale boat. Both experimental results from sea testing and numerical results from a simulator, which was designed for this project, are investigated and compared throughout the thesis period. The input of the external conditions that were acting on the boat are used for comparison in the simulations. The results reflect clear observations on the behaviour of such sailing model. Furthermore, the quality of the simulations can be assessed through the interpretation of the discrepancies present between results. This leads to improvements of the coefficients and assumptions used in the simulations. Regarding the conclusions of this master thesis, future testing and data recollection will improve the accuracy of observations.

performance analysis --- manoeuvrability --- zigzag --- turning circle --- hydrodynamics --- aerodynamics --- wind propulsion --- cargo shipping --- sailing --- marine research --- experimental testing --- Ingénierie, informatique & technologie > Ingénierie aérospatiale

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

This thesis is a contribution towards model-based guidance for the search of victims of drowning in rivers. It includes two parts: (i) an experimental study aiming at characterizing the hydrodynamic properties of a human-body shape, and (ii) the setup of a preliminary computational tool designed to simulate the drift of a human body in an open channel.
Literature on the prediction of the motion of floating objects and bodies in water was first reviewed. For the case of a human body, a lack of knowledge on the parameters influencing controlling the drift was highlighted.
Next, two complementary series of laboratory experiments were undertaken, one in a 20-m long glass flume and the second one in a 100-m long towing tank. The former involves rigid body models (dummies) at a scale of approximately 1:6, the covered range of relative velocity is 0.09 and 0.2 m/s and the particulate Reynolds number spans between 2.8*104 and 5.2*104. In the towing tank, a prototype-scale rigid body model was used, with a velocity of 0.1 to 0.5 m/s, and a particulate Reynolds number varying between 105 and 8*105.
The experiments enabled estimating the drag coefficient of the body models, and the results were compared to a reference shape (cylinder). A systematic analysis of parameters such as the body shape, its orientation with respect to the approaching flow, and its submersion depth was undertaken. Additional parameters were also explored, including the presence of hair and clothes.
Based on a combination of backwater profile calculation and parametrized cross-sectional velocity distribution, a 3D computational model of the flow field in the laboratory flume was set up. Simple drift computations were performed, which focused on the body motion in the streamwise direction. This enabled sensitivity analysis regarding the relative importance of the drag and added mass coefficients.
Finally, limitations of the study are discussed and future research directions are proposed.

drag --- body-like shape --- experiment --- drowning --- drift --- Ingénierie, informatique & technologie > Ingénierie civile

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Dans ce travail, une étude approfondie a été réalisée sur les tourbillons de coin qui sont observables sur le toit plat d'un bâtiment (phénomène illustré par Banks et al. (2000)). Les mesures de soufflerie utilisées pour ce travail proviennent de la thèse de Blaise (2016). Elles concernent 121 capteurs. Les tests en soufflerie ont été réalisés pendant 13 heures. Suite à ces observations, on a remarqué de larges sous-pressions qui vont fortement influencer le dimensionnement de la structure. Il a été possible d'observer que les densités de probabilité des coefficients de pression obtenus en soufflerie ont un comportement s'apparentant à un mélange de deux densités de probabilité. Une approche de dé-mixage a été traitée à l'Université de Liège par le passé par Bourcy (2020) et Rigo et al. (2020).
Dans ce travail, une étude générale a été réalisée pour des signaux ayant une densité de probabilité mélangée par deux densités de probabilité normales transformées par une cubique. Les développements mathématiques des mélanges ont été réalisés pour les densités de probabilité conjointes et tri-variées. Les formulations mathématiques utilisées peuvent facilement être transposées pour une densité de probabilité dépendante de n signaux mélangés. La théorie développée est générale et pourra être utilisée dans une multitude de domaines traitant des signaux mélangés. A l'aide des formulations mathématiques, il a été possible de réaliser un algorithme d'Exp-Max permettant de définir les statistiques unilatérales et croisées de chacune des composantes des mélanges obtenus pour les mesures en soufflerie. Une amélioration de l'algorithme Exp-Max a été utilisée afin de traiter des vecteurs de mesures de grande longueur.

À l'aide du nouveau modèle créé, on a pu réaliser différentes études sur le comportement des statistiques croisées. On a défini des zones avec des comportements différents, qui ont été étudiées séparément. Une étude a été faite sur le comportement spatial d'intégration de corrélation permettant d'améliorer la disposition des capteurs sur les modèles en soufflerie. On a aussi étudié le comportement des lignes d'intégration de corrélation et de covariance illustrant les relations spatiales entre capteurs. Cela permet de mettre en évidence les problèmes avec l'approche TVL et la nécessité de trouver un modèle plus performant dans le futur. L'approche actuelle TVL utilise un paramètre adimensionnel K. Il existe plusieurs valeurs pour le coefficient K. Le choix de la valeur de ce dernier va fortement influencer les résultats qui peuvent être transmis aux bureaux d'études.

Ensuite, une formulation mathématique a été développée de manière à calculer la moyenne et l'écart-type d'une résultante de pression pour un ensemble de n capteurs en considérant que les signaux ont une densité de probabilité mélangée par deux densités de probabilité normales transformées par une fonction cubique.

Au vu de la difficulté de reconstruction des phénomènes spatiaux à l'aide de mesures locales, un nouveau type de capteur a été imaginé en conclusion de ce travail. Une solution consisterait à ne plus travailler avec des capteurs de pression locaux, mais avec des capteurs de surface. Travailler de cette manière permettrait d'intégrer les effets spatiaux de manière physique sans passer par une formulation mathématique approchant la réalité.

Ingénierie, informatique & technologie > Ingénierie civile