This thesis investigates Stochastic Separated Flow (SSF) models for particle dispersion in turbulent shear flows. A new model is presented that accounts for anisotropy and incorporates a temporal and a spatial autocorrelation in the description of the fluctuating component of the turbulent gas-phase velocity. This model and three SSF models available in the literature are evaluated by comparing predictions with the shear layer experiments of Lazaro and Lasheras (1989), Hishida et al (1992) and the turbulent round jet experiment of Yuu et al (1978). Results are discussed and deficiencies in the models explored. The new model of Evinou and Lightstone compensates for the crossing trajectory effect with the inclusion of a spatial correlation based on the relative velocity of the particle and the time step employed. / Thesis / Master of Applied Science (MASc)

Engine development has been in the direction of increased turbine inlet temperatures to improve efficiency and power output. Secondary flows develop as a result of a near-wall pressure gradient in the stagnating flow approaching the inlet nozzle guide vane as well as a strong cross-passage gradient within the passage. These flow structures enhance heat transfer and convect hot core flow gases onto component surfaces. In modern engines it has become critical to cool component surfaces to extend part life. Bypass leakage flow emerging from the slot between the combustor and turbine endwalls can be utilized for cooling purposes if properly designed. This study examines a three-dimensional slot geometry, scalloped to manipulated leakage flow distribution. Statistical techniques are used to decouple the effects of four geometric parameters and quantify the relative influence of each on endwall cooling levels and near-wall total pressure losses. The slot geometry is also optimized for robustness across a range of inlet conditions. Average upstream distance to the slot is shown to dominate overall cooling levels with nominal slot width gaining influence at higher leakage flow rates. Scalloping amplitude is most influential to near-wall total pressure loss as formation of the horseshoe vortex and cross flow within the passage are affected. Scalloping phase alters local cooling levels as leakage injection is shifted laterally across the endwall. / Ph. D.

This thesis focuses on the behaviour of hyperpycnal plumes in river mouth discharges. The plunging of high density flows in two dimensional channels has been extensively studied before. A fundamental assumption in these studies is that the flow is laterally confined. These studies allow the flow to plunge only in two directions, the horizontal x-direction and the vertical z-direction. The goal of this study is to determine if there is observable plunging of hyperpycnal flows in the lateral y-direction, i.e. lateral spreading, in a three dimensional domain and to find out the parameters influencing the lateral spread. Previous studies conducted in laterally confined channels suggest that hyperpycnal flows plunge when the flow reaches a densimetric Froude number of unity. This study attempts to find the densimetric Froude number at hyperpycnal plunging in a three dimensional domain and if it is influenced by the factors that also influence the spread. This study also analyzes whether the cross-shore location for plunging changes when lateral spreading is accounted for, relative to a two dimensional analysis and if the plunging is limited to flow reaching a certain depth. This was accomplished through a series of experimental simulations on a hypothetical river mouth domain using Delft-3D, a hydrodynamic modeling software. Three parameters viz. the bottom slope of the receiving basin, the bottom friction and the density difference between inflow and ambient liquid were varied to test their influence on the plume spread rate. / Master of Science / It is crucial for researchers to have the expertise in modeling flow processes that develop in oceans, lakes and reservoirs in order to aid efforts in improving conditions for water quality within such domains. Hyperpycnal flows, also commonly known as high density flows are among one of the the less studied phenomenon in this discipline. This phenomenon occurs when a river carrying water with high density flows into an ocean, lake or a reservoir containing water with a lower density. Such flow regimes cause the inflow to submerge and sink to the bottom (plunge) and form a density current on the bed of the receiving basin. Studying density flows is important to model the transport of sediments, dissolved solids or pollutants. This study aims to improve the existing understanding of hyperpycnal plumes, their plunge location and spread in a three dimensional domain. For this, a simulation software Delft3D was used to build a model that is representative of the system and closely resembles the flow processes taking place in the aforementioned domains. Simulations were then run to collect data on how factors like the initial flow conditions (∆ρ), the basin slope (S) and friction (Chézy coefficient, Cz) have an impact on the phenomenon. This data was then compared to previous analyses to show the difference in plume behaviour and prediction of plunging. This study serves as a stepping stone in the ultimate goal of developing a prediction model for hyperpycnal plumes, indicating that Delft3D is a promising tool for analyzing such phenomenon.

When applying optimization techniques to complex physical systems, using very large numerical models for the solution of a system of parameter dependent partial differential equations (PDEs) is usually intractable. Surrogate models are used to provide an approximation to the high fidelity models while being computationally cheaper to evaluate. Typically, for time dependent nonlinear problems a reduced order model is built using a basis obtained through proper orthogonal decomposition (POD) and Galerkin projection of the system dynamics. In this thesis we present theoretical and numerical results for parameter dependent model reduction techniques. The methods are motivated by the need for surrogate models specifically designed for nonlinear parameter dependent systems. We focus on methods in which the projection basis also depends on the parameter through extrapolation and interpolation. Numerical examples involving 1D Burgers' equation, 2D Navier-Stokes equations and 2D Boussinesq equations are presented. For each model problem comparison to traditional POD reduced order models will also be presented. / Ph. D.

Cardiovascular diseases (CVD) are the leading cause of death in the world. According to the World Health Organization (WHO) 17.3 million people died from cardiovascular disease in 2008, representing 30% of all global deaths. The most common modality of treatment of occluded arteries is the use of stents. Despite the widespread use of stents, the incidence of post-stent restenosis is still high. The study of stents in conditions that are similar to in-vivo conditions is limited. This work tries to address the behavior of stents in conditions similar to in-vivo conditions in a generalized framework, thus providing insights for stent design and deployment. Three dimensional, time accurate computational fluid dynamics (CFD) simulations in a pulsatile flow with fluid-structure interaction (FSI) were carried out in realistic coronary arteries, with physiologically relevant flow parameters and dynamics due to induced motion of the heart. In addition, the geometric effects of the stent on the artery were studied to point towards possible beneficial stent deployment strategies. The results suggest that discontinuities in compliance and dynamic geometry cause critical changes in local hemodynamics, namely altering the local pressure and velocity gradients. Increasing the stent length, reducing the transition length and increasing the overexpansion caused adverse flow conditions. From this work, detailed flow characteristics and hemodynamic characteristics due to the compliance mismatch and applied motion were obtained that gave insights towards better stent design and deployment. / Ph. D.

The ultimate goal of this work is to facilitate the design of gas turbine engine particle separators by reducing the computational expense to accurately simulate the fluid flow and particle motion inside the separator. It has been well-documented that particle ingestion yields many detrimental impacts for gas turbine engines. The consequences of ice particle ingestion can range from surface-wear abrasion to engine power loss. It is known that sufficiently small particles, characterized by small particle response times (τp), closely follow the fluid trajectory whereas large particles deviate from the streamlines. Rather than manually deriving how the particle acceleration varies from the fluid acceleration, this work chooses to implicitly derive this relationship using machine learning (ML). Inertial particle separators are devices designed to remove particles from the engine intake flow, which contributes to both elongating the lifespan and promoting safer operation of aviation gas turbine engines. Complex flows, such as flow through a particle separator, naturally have rotation and strain present throughout the flow field. This study attempts to understand if the motion of particles within rotational and strained canonical flows can be accurately predicted using supervised ML. This report suggests that preprocessing the ML training data to the fluid streamline coordinates can improve model training. ML models were developed for predicting particle acceleration in laminar, fully rotational/irrotational flows and combined laminar flows with rotation and strain. Lastly, the ML model is applied to particle data extracted from a Computational Fluid Dynamics (CFD) study of particle-laden flow around a louver-geometry. However, the model trained with particle data from combined canonical flows fails to accurately predict particle accelerations in the CFD flow field. / Master of Science / Aviation gas turbine engine particle ingestion is known to reduce engine lifespans and even pose a threat to safe operation in the worst case. Particles being ingested into an engine can be modeled using multiphase flow techniques. Devices called inertial particle separators are designed to remove particles from the flow into the engine. One challenge with designing such a separator is figuring out how to efficiently expel the small particles from the flow while not unnecessarily increasing pressure loss with excessive twists and turns in the geometry. Designers usually have to develop such geometries using multiphase flow computational fluid dynamics (CFD) that solve the fluid and particle dynamics. The abundance of data associated with CFD, and especially multiphase flows make it an ideal application to study with machine learning (ML). Because such multiphase simulations are very computationally expensive, it is desirable to develop "cheaper" methods. This is the long term goal of this work; we want to create ML surrogates that decrease the computational cost of simulating the particle and fluid flow in particle separator geometries such that designs can be iterated more quickly. In this work we introduce how artificial neural networks (ANNs), which are a tool used in ML, can be used to predict particle acceleration in fluid flow. The ANNs are shown to learn the acceleration predictions with acceptable accuracy for the training data generated with canonical flow cases. However, the ML model struggles to become generalizable to actual CFD simulations.

We investigate in detail two multiuser opportunistic scheduling problems in centralized wireless systems: the scheduling of "delay-sensitive" flows with packet delay requirements of a few tens to few hundreds of milliseconds over the air interface, and the scheduling of "best-effort" flows with the objective of minimizing mean file transfer delay. Schedulers for delay-sensitive flows are characterized by a fundamental tradeoff between "maximizing total service rate by being opportunistic" and "balancing unequal queues (or delays) across users". In choosing how to realize this tradeoff in schedulers, our key premise is that "robustness" should be a primary design objective alongside performance. Different performance objectives -- mean packet delay, the tail of worst user's queue distribution, or that of the overall queue distribution -- result in remarkably different scheduling policies. Different design objectives and resulting schedulers are also not equally robust, which is important due to the uncertainty and variability in both the wireless environment and the traffic. The proposed class of schedulers offers low packet delays, less sensitivity to the scheduler parameters and channel characteristics, and a more graceful degradation of service in terms of the fraction of users meeting their delay requirements under transient overloads, when compared with other well-known schedulers. Schedulers for best-effort flows are characterized by a fundamental tradeoff between "maximizing the total service rate" and "prioritizing flows with short residual sizes". We characterize two regimes based on the "degree" of opportunistic gain present in the system. In the first regime -- where the opportunistic capacity of the system increases sharply with the number of users -- the use of residual flow-size information in scheduling will 'not' result in a significant reduction in flow-level delays. Whereas, in the second regime -- where the opportunistic capacity increases slowly with the number of users -- using flow-size information alongside channel state information 'may' result in a significant reduction. We then propose a class of schedulers which offers good performance in either regime, in terms of mean file transfer delays as well as probability of blocking for systems that enforce flow admission control. This thesis provides a comprehensive theoretical study of these fundamental tradeoffs for opportunistic schedulers, as well as an exploration of some of the practical ramifications to engineering wireless systems. / text

La synchronisation des flux physiques a reçu une importante attention dans la littérature. Si le non flux physique a reçu une attention toute particulière dans des domaines tels que le SCM, le pendant financier a longtemps était délaissé. Plus inquiétant, les délais de paiement ont longtemps été perçus comme un moyen de réduction du besoin en fonds de roulement des grandes entreprises. La crise des Subprimes a eu de lourdes conséquences sur le financement des entreprises les plus modestes, mettant en danger l’existence même de ces dernières. Notre travail initial de synchronisation des flux physiques et financiers instruits dans la cadre d’un partenariat CIFRE visait à trouver des solutions à ces problématiques en proposant des modèles d’optimisation sous contraintes. Néanmoins, notre volonté de mettre en œuvre nos travaux fut vaine et notre sujet a progressivement drifté vers la découverte de l’échec du déploiement d’un système ERP. Notre enracinement au sein d’une grande entreprise du secteur de la défense française nous a permis de mettre en œuvre une recherche action canonique en trois phases distinctes et a débouché sur la production d’un modèle de diffusion de l’innovation appliqué à l’ERP. Ce modèle en 6 phases, décrits les étapes successives du déploiement, en mettant en exergue l’impact des forces politiques au sein de l’organisation. / Synchronization of physical flows received significant attention in the literature. If the non-physical flows received special attention in areas such as SCM, financial for a long time was helpless. More worryingly, payment delays have long been seen as a means of reducing the need for working capital for large firms. The subprime crisis has had a serious impact on the financing of the smaller companies, endangering the very existence of the latter. Our initial synchronization job physical and financial flows educated in the context of a CIFRE partnership aimed at finding solutions to these problems by proposing models of optimization under constraints. However, our commitment to implement our work was in vain and our subject has gradually drifted to the discovery of the failure of the deployment of an ERP system. Our roots in a large enterprise sector French defense allowed us to implement an action research canonical three distinct phases and resulted in the production of a model of diffusion of innovation applied to the ERP. This model into 6 phases, described the successive stages of deployment, highlighting the impact of political forces within the organization

O objetivo principal desta dissertação é mostrar um resultado obtido por Plante, o qual estabelece que: qualquer fluxo de Anosov de codimensão um sobre uma variedade diferenciável compacta M de dimensão maior do que 3 com grupo fundamental solúvel é topologicamente equivalente à suspensão de um automorfismo hiperbólico do toro. Este resultado mostra a conjectura de Verjovsky no caso que o grupo fundamental da variedade é um grupo solúvel. A prova deste resultado é baseada no celebre resultado de Schwartzman, o qual fornece um criterio para garantir a existencia de seção transversal global para um fluxo não singular / O objetivo principal desta dissertação é mostrar um resultado obtido por Plante em [12] o qual estabelece que: qualquer fluxo de Anosov de codimensão um sobre uma variedade diferenciável compacta M de dimesão maior o que 3 com grupo fundamental solúvel é topologicamente equivalente à suspensão de um automorfismo hiperbólico do toro. Este resultado mostra a conjectura de Verjovsky no caso que o grupo fundamental da variedade é um grupo solúvel. A prova deste resultado é baseada no célebre resultado de Schwartzman [15], Teorema 2.17, o qual fornece um critério para garantir a existência de seção transversal global para um fluxo não singular

The purpose of this work is to study the effects of boundaries on granular flows down vibrating inclines, on segregation in granular mixtures induced by boundary vibrations, and on flows of granular mixtures through vibrating sieves. In each case, we employ techniques borrowed from the kinetic theory to derive an appropriate set of boundary conditions, and combine them with existing flow theories to calculate the profiles of solid volume fraction, mean velocity, and granular temperature throughout the flows. The boundaries vibrate with full three-dimensional anisotropy in a manner that can be related to their amplitudes, frequencies, and phase angles in three independent directions. At impenetrable surfaces (such as those on the inclines), the conditions derived ensure that momentum and energy are each balanced at the boundary. At penetrable surfaces (such as sieves), the conditions also ensure that mass is balanced at the boundary. In these cases, the momentum and energy balances also are modified to account for particle transport through the boundary. Particular interest in all the applications considered here is in how the details of the boundary geometry and the nature of its vibratory motion affect the resulting flows. In one case, we derive conditions that apply to a monosized granular material that interacts with a bumpy, vibrating, impenetrable boundary, and predict how such boundaries affect steady, fully developed unconfined inclined flows. Results indicate that the flows can be significantly enhanced by increasing the total energy of vibration and are more effectively enhanced by normal vibration than by tangential vibration. Regardless of the direction of vibration, the bumpiness of the boundary has a profound effect on the flows. In a second case, we derive conditions that apply to a binary granular mixture that interacts with a flat, vibrating, penetrable sieve-like boundary, and predict how such boundaries affect the process in which the particles pass through the sieve. In the special case in which the particles are all the same size, the results make clear that energy is more effectively transmitted to the assemblies when either the total vibrational energy or the normal component of the vibrational energy is increased, but that an increase in the energy transferred to the material can sometimes actually decrease the flow rates through the sieve. Consequently, at any instant of time in the sieving process, there is an optimum level of vibrational energy that will maximize the flow rate. For the sieving of binary granular assemblies, the physics associated with the effects of energy transfer on the flow rates still applies. However, in these cases, the flows through the sieve are also profoundly affected by segregation that occurs while the particles reside on sieve before the pass through. For this reason, we also isolate the segregation process from the sieving process by considering the special case in which the holes in the vibrating sieve are too small to allow any particles to pass through. In this case, the results show that under most circumstances the region immediately adjacent to the vibrating surface will be populated almost entirely by the smaller particles or by the more dissipative particles if there is no size disparity, and that the reverse is true in a second region above the first.