A fire growth calculation method has been developed that couples a computational fluid dynamics (CFD) model with bench scale cone calorimeter test data for predicting the rate of flame spread on compartment contents such as furniture. The commercial CFD code TASCflow has been applied to solve time averaged conservation equations using an algebraic multigrid solver with mass weighted skewed upstream differencing for advection. Closure models include k-epsilon for turbulence, eddy breakup for combustion following a single step irreversible reaction with Arrhenius rate constant, finite difference radiation transfer, and conjugate heat transfer. Radiation properties are determined from concentrations of soot, CO2 and H2O using the narrow band model of Grosshandler and exponential wide band curve fit model of Modak. The growth in pyrolyzing area is predicted by treating flame spread as a series of piloted ignitions based on coupled gas-fluid boundary conditions. The mass loss rate from a given surface element follows the bench scale test data for input to the combustion prediction. The fire growth model has been tested against foam-fabric mattresses and chairs burned in the furniture calorimeter. In general, agreement between model and experiment for peak heat release rate (HRR), time to peak HRR, and total energy lost is within pm 20%. Used as a proxy for the flame spread velocity, the slope of the HRR curve predicted by model agreed with experiment within pm 20% for all but one case.

Este trabalho tem como objetivo desenvolver ferramenta computacional para simulação e análise do fenômeno de penetração e cravação de estacas torpedos em solo marítimo. A abordagem será baseada no método Moving Particle Semi-Implicit (MPS). Por se tratar de um método de partícula, sem malha, o mesmo apresenta grande flexibilidade na modelagem de problemas de interação fluido-sólido com fragmentação ou junção de superfície livre e grandes deslocamentos ou deformações dos sólidos, fenômenos esses presentes no impacto e cravação da estaca no solo marítimo. Para isso, dois desafios foram elencados: a modelagem dos solos como fluidos não-newtonianos e a determinação da força de arrasto viscosa na superfície de sólidos. A modelagem do fluido não-newtoniano foi feita considerando os modelos de Power Law, Bingham e Herschel-Bulkley. O cálculo da força de arrasto viscosa foi avaliado determinando-se o gradiente da velocidade do fluido na direção normal à parede com base na regressão polinomial. Por simplicidade, foi considerada a hipótese de que a variação da velocidade na direção tangencial da parede é muito menor se comparada a variação da mesma na direção do vetor normal. O método implementado, assim como o escoamento de fluidos não-newtonianos, foi validado por meio de comparação entre o resultado obtido de simulações com geometrias pré-definidas e as respostas analíticas para tais casos. Como exemplo de aplicação da ferramenta computacional desenvolvida, um caso simplificado de cravação das estacas torpedos foi simulado avaliando-se o seu deslocamento dentro do solo e os esforços cisalhantes a ela submetidas. / This work aims to develop computational tools to simulate and analysis the torpedo anchor penetration in marine soil. The approach will be based on the Moving Particle Semi-Implicit (MPS) method. Because it is a meshless method, it is extremely flexible to model fluid-solid interaction with fragmentation or junction of free surface and large displacements or deformations of solids, phenomena presented at the torpedo anchor impact. Two challenges were listed: the modeling of soils as non-Newtonian fluids and the determination of the viscous drag on the solids surface. The modeling of non-Newtonian fluid was done based on the Power Law, Bingham and Herschel-Bulkley models. The calculation of the viscous drag was evaluated by determining the velocity gradient in the normal direction of the wall based on polynomial regression considering the fluid particles near the solid wall. In this work, for sake of simplicity, the hypothesis that the velocity variation in the tangential direction of the wall is much smaller compared to its variation in the normal direction is adopted. The proposed technique, as well as the flow of non-Newtonian fluids, were validated comparing the results obtained in flow simulations with predefined geometries with the expected analytical responses. As an example of the application of the computational tools developed, a simplified case of torpedo penetration was simulated by evaluating its displacement and the shear stresses submitted to it.

In this thesis, a numerical method is developed for simulating non-isothermal multiphase flows, which are important in many technical applications such as crystal growth and welding. The method is based on the arbitrary Lagrangian Eulerian method of Li (2013). The interface is represented explicitly by mesh lines, and is tracked by an adaptive moving unstructured mesh. The $P2-P1d$ finite element method (FEM) is used for discretisation and the incompressible Navier-Stokes equations are solved by the uzawa method. Firstly, a thorough study is presented on the method's capability in numerically representing the force balance condition on the interface. An inaccurate representation of this condition induces the non-physical spurious currents, which degrade the simulation accuracy especially when the viscous damping is weak (small Ohnesorge number, $Oh$). For the example of a circular/spherical droplet, the interfacial tension and the associated pressure jump are exactly balanced numerically and thus the static Laplace solution exists in our method. The stability of this solution is examined numerically. The amplitude of the dimensionless spurious currents is found to be around $10^{−15}$ for $Oh \geq 10^{−3} $. Another benchmark test is the axisymmetric oscillation of a freesurface droplet/bubble. The simulation results are in good agreement with the analytical solution for $Oh = 10^{−3}$. This is by far the first successful simulation of droplet/bubble oscillation with such weak viscous damping and it demonstrates the ability of our method in simulating flows with strong capillary forces. Secondly, a numerical treatment of interface topology changes is incorporated into our method for studying problems with interface breakup. Thanks to the adaptive mesh generator, the thin region between the interface boundary and another boundary consists of one layer of elements. The interface topology change is performed once the minimum distance between the two boundaries falls below a pre-set scale $l_{breakup}$ . The numerical implementation is verified through two different examples: dripping faucet and droplet coalescence. Remarkably good agreement has been obtained with the experimental results. The simulation of the low Oh dripping problem shows both the accuracy and robustness of our method. The simulation of droplet coalescence demonstrates the great advantage of our method in solving problems with a large disparity in length scales. Finally, an FEM solver for temperature is developed and the non-isothermal effects are included in our method for the purpose of simulating non-isothermal multiphase flows. The modified method is validated to be accurate through three benchmark examples: natural convection in a cavity, thermocapillary convection of two layers, and droplet migration subject to a temperature gradient. Our method is then applied to investigate the liquid bridge breakup with thermocapillary effect. The non-isothermal liquid bridge breakup in the viscous and inertial regimes are studied. It has been found that the inertial regime breakup exhibits different pinchoff shapes as the Capillary number increases, and that the viscous regime breakup is accelerated by the thermocapillary motion.

L’industrie automobile se trouve aucentre des préoccupations environnementalesactuelles. Les moteurs essence à injection directeopérés en condition pauvres offrent un fortpotentiel en terme de réduction des émissions depolluants. En contrepartie, ils sont sujets à uneforte variabilité cyclique de combustion (CCV)qui ne peut être que partiellement étudiéeexpérimentalement. La simulation aux grandeséchelles (SGE) apparait comme une approchenumérique adaptée pour étudier de telsphénomènes, du fait de sa capacité naturelle àcapter les phénomènes instationnaires. Laprésente thèse se propose d’une part d’estimer lacapacité de la SGE à reproduire les CCVobservées expérimentalement, et d’autre part decontribuer à une meilleure compréhension deleur apparition. Dans ce contexte, un effortparticulier est mis sur la modélisation desphénomènes proche paroi. En particulier, unmodèle de paroi adapté à l’étude del’aérodynamique interne de configurationindustrielles est proposé. Il est validé sur desconfigurations de complexités variées. Enfin,l’étude porte sur le cas du moteur M256 qui estétudié en s’appuyant sur une solide base dedonnées expérimentales. Les causes de CCV sontexplorées, et notamment, l’impact de lavariabilité de l’écoulement généré pendant laphase d’admission sur la propagation du front deflamme est clarifié. / The automotive industry finds itselfat the center of current environmental concerns.Modern direct injection engines, operated underlean condition have the potential to reducepollutant emissions. As a drawback, they aresubject to large cyclic combustion variability(CCV), that can be explained only partially byexperimental measurements. Large-EddySimulation (LES) appears as an adapted tool tocomplement experiments, due to its naturalability to capture unsteady phenomena. Thepresent PhD thesis first aims at reproducing theCCV, and at contributing achieving a betterunderstanding of their occurrence.In this context, a special effort is put on thereproduction of near-wall phenomena, throughthe proposal of a new wall boundary conditionthat is validated on cases of differentcomplexity. Then the focus is put on the M256case, for which an extensive experimentaldatabase is available. The causes of CCV areexplored, and in particular, the impact of thevariability of the intake flow on the flame frontpropagation is clarified.

This PhD involved computational fluid dynamic simulations of finned generators cooling under forced convection in an oscillating water column environment. Various design changes to the upstream Wells turbine and its effect on the consequent cooling of the generator were investigated. Simulations were run in steady-state to obtain an initial condition, thereafter, unsteady simulations revealed a steadying of heat transfer over the course of multiple blade rotation cycles. This justified the use of steady-state for the remaining simulations over a range of flow coefficients. The results revealed that the heat transfer from the generator increased for tighter blade tip clearances, thicker blade profiles and greater turbine solidity. The heat transfer was found to increase with rising flow rate coefficient, which was adjusted by increasing the inlet velocity whilst maintaining the angular velocity of the turbine at a constant 2000 RPM. Additionally, the variation of turbine angular velocity at a fixed flow rate coefficient was investigated, the heat transfer was also found to increase with angular velocity, albeit by a far lesser extent. The inclusion of the Wells turbine upstream of the generator was investigated initially and was found to increase heat transfer due to the resulting impingement of airflow across the generator. In all design scenarios in which the heat transfer increases, there is also an observed increase in the mass flow rate of air, radially, towards the generator.

Les satellites sont mis en orbite en utilisant des lanceurs dont la conception est une des activités principales d’Airbus Defence and Space. Pour ce faire, se baser sur des expériences n’est pas facile : les souffleries ne permettent pas d’évaluer toutes les situations auxquelles un lanceur est confronté au cours de sa mission. La simulation numérique est donc essentielle pour l’industrie spatiale. Afin de disposer de simulations toujours plus fidèles, il est nécessaire d’utiliser des supercalculateurs de plus en plus puissants. Cependant, ces machines voient leur complexité augmenter et pour pouvoir exploiter leur plein potentiel, il est nécessaire d’adapter les codes existants. Désormais, il semble essentiel de passer par des couches d’abstraction afin d’assurer une bonne portabilité des performances. ADS a développé depuis plus de 20 ans le code FLUSEPA qui est utilisé pour le calcul de phénomènes instationnaires comme les calculs d’onde de souffle au décollage ou les séparations d’étages. Le solveur aérodynamique est basé sur une formulation volume fini et une technique d’intégration temporelle adaptative. Les corps en mouvement sont pris en compte via l’utilisation de plusieurs maillages qui sont combinés par intersections.Cette thèse porte sur la parallélisation du code FLUSEPA. Au début de la thèse, la seule version parallèle disponible était en mémoire partagée. Une première version parallèle en mémoire distribuée a d’abord été réalisée. Les gains en performance de cette version ont été évalués via l’utilisation de deux cas tests industriels. Un démonstrateur du solveur aérodynamique utilisant la programmation par tâche au dessus d’un runtime a aussi été réalisé. / There are different kinds of satellites that offer different services like communication, navigationor observation. They are put into orbit through the use of launchers whose design is oneof the main activities of Airbus Defence and Space. Relying on experiments is not easy : windtunnel cannot be used to evaluate every critical situation that a launcher will face during itsmission. Numerical simulation is therefore mandatory for spatial industry.In order to have more reliable simulations, more computational power is needed and supercomputersare used. Those supercomputers become more and more complex and this impliesto adapt existing codes to make them run efficiently. Nowadays, it seems important to rely onabstractions in order to ensure a good portability of performance. Airbus Defence and Spacedeveloped for more than 20 years the FLUSEPA code which is used to compute unsteady phenomenalike take-off blast wave or stage separation. The aerodynamic solver relies on a finitevolume formulation and an explicit temporal adaptive solver. Bodies in relative motion are takeninto account through the use of multiple meshes that are overlapped.This thesis is about the parallelization of the FLUSEPA code. At the start of the thesis,the only parallel version available was in shared memory through OpenMP. A first distributedmemory version was realized and relies on MPI and OpenMP. The performance improvementof this version was evaluated on two industrial test cases. A task-based demonstrator of theaerodynamic solver was also realized over a runtime system.

Wave energy is one of the most promising renewable energy sources for future exploitation. This thesis focuses on thermodynamic effects within Oscillating Water Column (OWC) devices equipped withWells turbines, particularly humidity effects. Previous theoretical studies of the operation of OWCs have resulted in expressions for the oscillation of the water surface in the chamber of an OWC based on linear wave theory, and the air expansion{compression cycle inside the air chamber based on ideal gas theory. Although in practice high humidity levels occur in OWC devices open to the sea, the influence of atmospheric conditions such as temperature and moisture on the performance of Wells turbines has not yet been studied in the field of ocean energy. Researchers have reported substantial differences between predicted and measured power output, and performance rates of OWCs presently coming into operation. The effect of moisture in the air chamber of the OWC causes variations on the atmospheric conditions near the turbine, modifying its performance and efficiency. Discrepancies in available power to the turbine are believed to be due to the humid air conditions, which had not been modelled previously. This thesis presents a study of the influence of humid air on the performance of an idealised Wells turbine in the chamber of an OWC using a real gas model. A new formulation is presented, including a modified adiabatic index, and subsequent modified thermodynamic state variables such as enthalpy, entropy and specific heat. The formulation is validated against experimental data, and found to exhibit better agreement than the ideal approach. The analysis indicates that the real gas behaviour can be explained by a non{dimensional number which depends on the local pressure and temperature in the OWC chamber. A first approach to the OWC formulation through the calculation of real air flow in the OWC is given, which predicts a 6% decrease in efficiency with respect to the ideal case when it is tested with a hypothetical pulse of pressure. This is important because accurate prediction of efficiency is essential for the optimal design and management of OWC converters. A numerical model has also been developed using computational fluid dynamics (CFD) to simulate the OWC characteristics in open sea. The performance of an OWC turbine is studied through the implementation of an actuator disk model in Fluent®. A set of different regular wave tests is developed in a 2D numerical wave flume. The model is tested using information obtained from experimental tests on a Wells{type turbine located in a wind tunnel. Linear response is achieved in terms of pressure drop and air flow in all cases, proving effectively the applicability of the actuator disk model to OWC devices. The numerical model is applied first to an OWC chamber containing dry air, and then to an OWC chamber containing humid air. Results from both cases are compared, and it is found that the results are sensitive to the degree of humidity of the air. Power decreases when humidity increases. Finally, results from the analytical real gas and numerical ideal gas models are compared. Very satisfactory agreement is obtained between the analytical and the numerical models when humidity is inserted in the gaseous phase. Both analytical and numerical models with humid air show considerable differences with the numerical model when dry air is considered. However, at the resonance frequency, results are independent of the gas model used. At every other frequency analysed, the real gas model predicts reduced values of power that can fall to 50% of the ideal power value when coupled to the radiation-diffraction model for regular waves. It is recommended that real gas should be considered in future analyses of Wells turbines in order to calculate accurately the efficiency and expected power of OWC devices.

Many highly developed physical models poorly approximate actual physical systems due to natural random noise. For example, convection in the earth's mantle—a fundamental process for understanding the geochemical makeup of the earth's crust and the geologic history of the earth—exhibits chaotic behavior, so it is difficult to model accurately. In addition, it is impossible to directly measure temperature and fluid viscosity in the mantle, and any indirect measurements are not guaranteed to be highly accurate. Over the last 50 years, mathematicians have developed a rigorous framework for reconciling noisy observations with reasonable physical models, a technique called data assimilation. We apply data assimilation to the problem of mantle convection with the infinite-Prandtl Boussinesq approximation to the Navier-Stokes equations as the model, providing rigorous conditions that guarantee synchronization between the observational system and the model. We validate these rigorous results through numerical simulations powered by a flexible new Python package, Dedalus. This methodology, including the simulation and post-processing code, may be generalized to many other systems. The numerical simulations show that the rigorous synchronization conditions are not sharp; that is, synchronization may occur even when the conditions are not met. These simulations also cast some light on the true relationships between the system parameters that are required in order to achieve synchronization. To conclude, we conduct experiments for two closely related data assimilation problems to further demonstrate the limitations of the rigorous results and to test the flexibility of data assimilation for mantle-like systems.

This thesis focuses on providing the means to use vertical-axis wind turbines (VAWTs) in wind farms as an alternative form of harnessing wind energy in offshore and urban environments where both wake and acoustic effects of turbines are important considerations. In order for VAWTs to be used in wind farm layout analysis and optimization, a reduced-order wake model is needed to calculate velocities around a turbine quickly and accurately. However, a VAWT wake model has not been available to accomplish this task. Using vorticity data from computational fluid dynamic (CFD) simulations of VAWTs and cross-validated Gaussian distribution and polynomial surface fitting, a wake model is produced that can estimate a wake velocity deficit of an isolated VAWT at any downstream and lateral position based on nondimensional parameters describing the turbine speed and geometry. When compared to CFD, which takes over a day to run one simulation, the wake model predicts the velocity deficit at any location with a normalized root mean squared error of 0.059 in about 0.02 seconds. The model agrees with two experimental VAWT wake studies with a percent difference of the maximum wake deficit of 6.3% and 14.6%. Using the actuator cylinder model with predicted wake velocities of multiple turbines, aerodynamic loads can be calculated on the turbine blades to estimate the power production of a VAWT wind farm. As VAWTs could be used in urban environments near residential areas, the noise disturbance coming from the turbine blades is an important consideration in the layout of a wind farm. Noise restrictions may be imposed on a wind farm to limit the disturbance, often impacting the wind farm's power producing capability. Two specific horizontal-axis wind turbine farm designs are studied and optimized using the FLORIS wake model and an acoustic model based on semi-empirical turbine noise calculations to demonstrate the impact a noise level constraint has on maximizing wind farm power production. When a noise level constraint was not active, the average power production increased, up to 8.01% in one wind farm and 3.63% in the other. Including a noise restriction in the optimization had about a 5% impact on the optimal average power production over a 5 decibel range. By analyzing power and noise together, the multi-modality of the optimization problem can be used to find solutions were noise impact can be improved while still maximizing wind farm power production.

Most fighters use the convergent-divergent nozzle configuration to accelerate into the supersonic realm. This nozzle configuration greatly increases the thrust potential of the aircraft compared to the simpler convergent nozzle. The nozzle design is not only crucial for thrust, but also for the drag since the afterbody drag can be as high as 15% of the total. Engine manufacturers optimize the engine and the nozzle configurations for the uninstalled conditions, but these may not be optimal when the engine is installed in the aircraft. The purpose of this study is to develop a methodology to optimize axisymmetric nozzle settings in order to maximize the net thrust. This was accomplished by combining both simulations of thrust and drag. The thrust model was created in an engine performance tool, called EVA, with the installed engine performance of a low bypass turbofan jet engine at maximum afterburner power setting. The drag model was created with CFD, where the mesh was built in ICEM Mesh and the simulations were run with the CFD solver M-Edge. Five Mach numbers in the range from 0.6 to 1.6 were simulated at an altitude of 12 km. The results showed that the afterbody drag generally decreased when increasing jet pressure ratio at both subsonic and supersonic velocities. At subsonic conditions, increasing nozzle area ratio for underexpanded nozzles would decrease the drag. Increasing nozzle area ratio for fully expanded or overexpanded nozzles would instead increase the drag to an intermediate point from where it would decrease. At supersonic condition, increasing nozzle area ratio would generally cause reduction in drag for all cases. The optimization showed that a net thrust increase of 0.02% to 0.09% could be gained for subsonic conditions while the supersonic optimization had negligible gain in thrust.