O presente estudo consistiu em avaliar os parâmetros aerodinâmicos de um modelo de ônibus rodoviário, comparando os resultados obtidos de simulação computacional via CFD (Computational Fluid Dynamics) com aqueles obtidos na experimentação em túnel de vento. O ônibus estudado foi do tipo rodoviário de um fabricante local, modelo Paradiso 1200. O veículo foi modelado em um software CAD (SolidWorks®) em duas escalas: 1/42 e 1/24. Além disso, para obter a comparação com a literatura, foram analisados dois tamanhos diferentes de um modelo do corpo de Ahmed. Posteriormente, foram criadas as malhas com as geometrias 3D e realizados os testes computacionais no software ANSYS FLUENT® para os quatro modelos, com o intuito de identificar alguns parâmetros aerodinâmicos como o coeficiente de arrasto, coeficiente de pressão, entre outros. Para as análises com o corpo de Ahmed foram utilizados os modelos de turbulência Spalart – Allmaras, κ – ε Standard, κ – ε RNG, κ – ω Standard, κ – ω SST e SST. Para os modelos de ônibus foram simulados apenas o modelo κ – ε Standard. Para a realização dos experimentos foi empregado um túnel de vento de circuito aberto, onde foram realizados testes de distribuição de pressão e arrasto aerodinâmico, variando a altura do vão livre entre a mesa automobilística e a superfície inferior dos modelos. Nos ensaios dos modelos onde houve a variação da altura em relação à mesa automobilística, foi identificado um aumento de 4,5% no valor do coeficiente de arrasto (Cd) para o corpo de Ahmed menor e 6,1% para o ônibus em escala 1/42. Comparando-se os resultados obtidos nos ensaios experimentais com aqueles obtidos nas análises numéricas, também ocorreram variações no Cd para todos os modelos. Nos ensaios de pressão o coeficiente de pressão (Cp) foi praticamente o mesmo entre os valores obtidos na análise em CFD e os valores experimentais, para ambos os modelos. Foram também realizados ensaios de visualização usando tufts de lã distribuídos na superfície externa do modelo menor de Ahmed e do modelo maior do ônibus. Esses ensaios indicaram nitidamente as regiões de recirculação de ar nos modelos, o que em parte não foi possível observar na análise computacional. Diante disso, verifica-se que os resultados experimentais obtidos em túnel de vento ainda são os mais confiáveis e utilizados, apesar dos altos custos envolvidos na construção de modelos, na instrumentação de alta tecnologia hoje disponível, nos métodos de visualização e na energia consumida nos testes. / Submitted by Ana Guimarães Pereira (agpereir@ucs.br) on 2016-12-09T18:14:15Z No. of bitstreams: 1 Dissertacao Giovanni Matheus Rech.pdf: 20273449 bytes, checksum: 70c97b02eb18e1d8b69454f3e92fe23d (MD5) / Made available in DSpace on 2016-12-09T18:14:15Z (GMT). No. of bitstreams: 1 Dissertacao Giovanni Matheus Rech.pdf: 20273449 bytes, checksum: 70c97b02eb18e1d8b69454f3e92fe23d (MD5) Previous issue date: 2016-12-09 / The present study was to evaluate the aerodynamic parameters of a road bus model by comparing the results of computer simulation via CFD (Computational Fluid Dynamics) with those obtained in experiments in a wind tunnel. The bus studied was a road type from a local manufacturer, Paradiso 1200 model. The vehicle was modeled on a CAD software (SolidWorks®) on two scales: 1/42 and 1/24. Furthermore, for comparison with the literature, we analyzed two different sizes of Ahmed body model. Thereafter, the meshes were created from 3D geometry and the computational tests performed with FLUENT® ANSYS software for the four models in order to identify some aerodynamic parameters such as the drag coefficient, pressure coefficient, among others. For analysis of Ahmed bodies, the turbulence models Spalart - Allmaras, κ - ε Standard, κ - ε RNG, κ - ω Standard, κ - ω SST and SST were used. For bus models, the turbulence model κ - ε Standard was only used. For the experiments we used an open circuit wind tunnel, where tests of pressure distribution and aerodynamic drag were performed, varying the height of the clearance between the automotive table and the bottom surface of the models. In the model tests, in which there were the height variation relative to the automotive table, an increase of 4.5% in the value of the drag coefficient (Cd) for the lower Ahmed body, and 6.1% for the bus 1/42 scale were identified. In pressure tests, the pressure coefficients (Cp) were almost the same between the values obtained from the CFD analysis and experimental values for both models. Visualization tests using wool tufts distributed on the outer surface of the smaller Ahmed model and the higher bus model were also performed. These tests clearly indicated the air recirculation regions in models, which in part was not observed in the computational analysis. Thus, it appears that the experimental results are in wind tunnel still the most reliable and used despite the high costs involved in the building models, in the high-tech instrumentation available today, in the visualization methods and in the energy consumed in the tests.

O presente estudo consistiu em avaliar os parâmetros aerodinâmicos de um modelo de ônibus rodoviário, comparando os resultados obtidos de simulação computacional via CFD (Computational Fluid Dynamics) com aqueles obtidos na experimentação em túnel de vento. O ônibus estudado foi do tipo rodoviário de um fabricante local, modelo Paradiso 1200. O veículo foi modelado em um software CAD (SolidWorks®) em duas escalas: 1/42 e 1/24. Além disso, para obter a comparação com a literatura, foram analisados dois tamanhos diferentes de um modelo do corpo de Ahmed. Posteriormente, foram criadas as malhas com as geometrias 3D e realizados os testes computacionais no software ANSYS FLUENT® para os quatro modelos, com o intuito de identificar alguns parâmetros aerodinâmicos como o coeficiente de arrasto, coeficiente de pressão, entre outros. Para as análises com o corpo de Ahmed foram utilizados os modelos de turbulência Spalart – Allmaras, κ – ε Standard, κ – ε RNG, κ – ω Standard, κ – ω SST e SST. Para os modelos de ônibus foram simulados apenas o modelo κ – ε Standard. Para a realização dos experimentos foi empregado um túnel de vento de circuito aberto, onde foram realizados testes de distribuição de pressão e arrasto aerodinâmico, variando a altura do vão livre entre a mesa automobilística e a superfície inferior dos modelos. Nos ensaios dos modelos onde houve a variação da altura em relação à mesa automobilística, foi identificado um aumento de 4,5% no valor do coeficiente de arrasto (Cd) para o corpo de Ahmed menor e 6,1% para o ônibus em escala 1/42. Comparando-se os resultados obtidos nos ensaios experimentais com aqueles obtidos nas análises numéricas, também ocorreram variações no Cd para todos os modelos. Nos ensaios de pressão o coeficiente de pressão (Cp) foi praticamente o mesmo entre os valores obtidos na análise em CFD e os valores experimentais, para ambos os modelos. Foram também realizados ensaios de visualização usando tufts de lã distribuídos na superfície externa do modelo menor de Ahmed e do modelo maior do ônibus. Esses ensaios indicaram nitidamente as regiões de recirculação de ar nos modelos, o que em parte não foi possível observar na análise computacional. Diante disso, verifica-se que os resultados experimentais obtidos em túnel de vento ainda são os mais confiáveis e utilizados, apesar dos altos custos envolvidos na construção de modelos, na instrumentação de alta tecnologia hoje disponível, nos métodos de visualização e na energia consumida nos testes. / The present study was to evaluate the aerodynamic parameters of a road bus model by comparing the results of computer simulation via CFD (Computational Fluid Dynamics) with those obtained in experiments in a wind tunnel. The bus studied was a road type from a local manufacturer, Paradiso 1200 model. The vehicle was modeled on a CAD software (SolidWorks®) on two scales: 1/42 and 1/24. Furthermore, for comparison with the literature, we analyzed two different sizes of Ahmed body model. Thereafter, the meshes were created from 3D geometry and the computational tests performed with FLUENT® ANSYS software for the four models in order to identify some aerodynamic parameters such as the drag coefficient, pressure coefficient, among others. For analysis of Ahmed bodies, the turbulence models Spalart - Allmaras, κ - ε Standard, κ - ε RNG, κ - ω Standard, κ - ω SST and SST were used. For bus models, the turbulence model κ - ε Standard was only used. For the experiments we used an open circuit wind tunnel, where tests of pressure distribution and aerodynamic drag were performed, varying the height of the clearance between the automotive table and the bottom surface of the models. In the model tests, in which there were the height variation relative to the automotive table, an increase of 4.5% in the value of the drag coefficient (Cd) for the lower Ahmed body, and 6.1% for the bus 1/42 scale were identified. In pressure tests, the pressure coefficients (Cp) were almost the same between the values obtained from the CFD analysis and experimental values for both models. Visualization tests using wool tufts distributed on the outer surface of the smaller Ahmed model and the higher bus model were also performed. These tests clearly indicated the air recirculation regions in models, which in part was not observed in the computational analysis. Thus, it appears that the experimental results are in wind tunnel still the most reliable and used despite the high costs involved in the building models, in the high-tech instrumentation available today, in the visualization methods and in the energy consumed in the tests.

Dans cette dissertation, nous considérons l’écoulement des liquides conducteurs d’électricité dans un champ magnétique externe. De tels écoulements sont décrits par les équations de la magnétohydrodynamique (MHD) quasi-statique, et sont fréquemment rencontrés dans des applications pratiques. Il suit qu’il y a un intérêt fort pour des outils numérques qui peuvent simuler ces écoulements dans des géometries complexes.<p>La première partie de cette thèse (chapitres 2 et 3) est dédiée à la présentation de la machinerie numérique qui a été utilisée et implémentée afin de résoudre les équations de la MHD quasi-statistique (incompressible). Plus précisément, nous avons contribué au développement d’un solveur volumes finis non-structuré parallèle. La discussion sur ces méthodes est accompagnée d’une analyse numérique qui est aussi valable pour des mailles non-structurées. Dans le chapitre 3, nous vérifions notre implémentation par la simulation d’un certain nombre de cas tests avec un accent sur des écoulements dans un champ magnétique intense.<p>Dans la deuxième partie de cette thèse (chapitres 4-6), nous avons utilsé ce solveur pour étudier des écoulements MHD de proche paroi .La première géometrie considérée (chapitre 4) est celle d’une conduite circulaire infini d’axe à haut nombre de Hartmann. Nous avons investitgué la sensitivité des résultats numériques au schéma de discrétisation et à la topologie de la maille. Nos résultats permettent de caractériser in extenso l’écoulement MHD dans une conduite avec des bords bien conducteurs par moyen des lois d’échelle.<p>Le sujet du cinquième chapitre est l’écoulement dans une conduite toroïdale à section carée. Une étude du régime laminaire confirme une analyse asymptotique pour ce qui concerne les couches de cisaillement. Nous avons aussi effectué des simulations des écoulements turbulents afin d’évaluer l’effet d’un champ magnétique externe sur l’état des couches limites limites.<p>Finalement, dans le chapitre 6, nous investiguons l’écoulement MHD et dans un U-bend et dans un coude arrière. Nous expliquons comment générer une maille qui permet de toutes les couches de cisaillement à un coût computationelle acceptable. Nous comparons nos résultats aux solutions asymptotiques. / Doctorat en Sciences / info:eu-repo/semantics/nonPublished

<p>Intra-Arterial Chemotherapy (IAC) is a preferred treatment for the primary liver cancer, despite its adverse side-effects. During IAC, a mixture of chemotherapeutic drugs, e.g. Doxorubicin, is injected into an artery supplying the tumor. A fraction of Doxorubicin is absorbed by the tumor, but the remaining drug passes into systemic circulation, causing irreversible heart failure. The efficiency and safety of the IAC can be improved by chemical filtration of the excessive drugs with a catheter-based Chemofilter device, as proposed by a team of neuroradilogists. </p> <p>The objective of my work was to optimize the hemodynamic and drug binding performance of the Chemofilter device, using Computational Fluid Dynamics (CFD) modeling. For this, I investigated the performance of two distinct Chemofilter configurations: 1) a porous “Chemofilter basket” formed by a lattice of micro-cells and 2) a non-porous “honeycomb Chemofilter” consisting of parallel hexagonal channels. A multiscale modeling approach was developed to resolve the flow through a representative section of the porous membrane and subsequently characterize the overall performance of the device. A heat and mass transfer analogy was utilized to facilitate the comparison of alternative honeycomb configurations. </p> A multiphysics approach was developed for modeling the electrochemical binding of Doxorubicin to the anionic surface of the Chemofilter. An effective diffusion coefficient was derived based on dilute and concentrated solution theory, to account for the induced migration of ions. Computational predictions were supported by results of <i>in-vivo</i> studies performed by collaborators. CFD models showed that the honeycomb Chemofilter is the most advantageous configuration with 66.8% drug elimination and 2.9 mm-Hg pressure drop across the device. Another facet of the Chemofilter project was its surface design with shark-skin inspired texturing, which improves the binding performance by up to 3.5%. Computational modeling enables optimization of the chemofiltration device, thus allowing the increase of drug dose while reducing systemic toxicity of IAC.

Den pågående elektrifieringen av transport och samhälle kräver utveckling av nya metoder för återvinning av batterier. Hydrometallurgi som fokuserar på selektiv kristallisation av metaller är ett intressant alternativ för dessa ändamål. Dessa system kan studeras genom modellering och simulering. Många matematiska modeller finns tillgängliga för att beskriva de olika involverade processerna i kristallisationen av metaller. Dessa processer inkluderar övermättnad, nukleation, kristalltillväxt och aggregation. Denna rapport sammanställer ett antal av de tillgängliga matematiska modellerna och presenterar ett numeriskt tillvägagångssätt för modellering av den tidsberoende nummerdensiteten av partiklar genom en populationsbalansekvation. Populationsbalansen kan lösas med olika metoder såsom momentmetoden och metoden av viktade residualer. Här löses ekvationen genom diskretisering. Diskretisering av den inre koordinaten i ett flertal längdintervall möjliggör simulering av partikel-storleksfördelningen som en funktion av tid. Det numeriska tillvägagångssättet applicerades på bariumsulfatutfällning i en perfekt blandad satsreaktor och två- och tre-dimensionella T-mixer-system, såväl som en perfekt blandad satsreaktor för förträngningskristallisation av nickelsulfat med groddning. Den simulerade storleksfördelningens placering visade sig ha bra överenstämmelse med experimentell data vid låga Reynolds-tal. Här undersöktes även påverkan av en mängd parametrar såsom diskretisering, aggregation och magnituden av diffusion. Aggregation hade en märkbar inverkan på välblandade system. Inverkan av aggregation i diffusions-kontrollerade system med kort retentionstid var låg. Diffusionsmagnituden hade liten påverkan på den normaliserade distributionen men större på det totala antalet partiklar. / The currently ongoing electrification of society and transport necessitates the development of novel methods for battery recycling. Hydrometallurgy with a focus on selective metal crystallisation is an interesting prospect to these ends. The resource recovery systems of interest can be studied through simulation where many mathematical models are available to describe the varying processes involved. These processes include supersaturation generation, nucleation, growth and aggregation. This work compiles some of these mathematical models and presents a numerical approach for the modelling of the time-dependent particle number density with a population balance equation. The population balance equation can be solved using a variety of different methods such as method of moments and method of weighted residuals. Here, the balance equation was solved by discretisation. Discretising the inner coordinate (crystal length) into a number of length intervals allows for the particle size distribution to be modelled as a function of time for various crystallisation systems. The framework was successfully applied to barium sulphate precipitation in a perfectly mixed batch reactor and two- and three-dimensional T-mixer systems, as well as a seeded perfectly mixed nickel sulphate anti-solvent crystallisation system. The simulated size distribution showed promising similarity to experimental data at low Reynolds number. The influence of a variety of parameters such as aggregation and magnitude of diffusion was investigated. Aggregation had a significant impact on well-mixed systems increasing with retention time. The impact of aggregation on diffusion-controlled systems with low retention time was low. The magnitude of diffusion had little impact on the particle size distribution of the crystal population but a large impact on the total number of crystals.

<p dir="ltr">On many hypersonic vehicles, gaps are present on the outer surface of the vehicle and the interaction of the hypersonic freestream flow over these gaps can cause significant heat transfer to the vehicle. The project described in this thesis analyzed selected hypersonic gap problems and attempted to offer solutions to combat the heat transfer occurring in the gap. The first section of this thesis is a parametric study to understand the changes to the heat transfer and flow that modifications to the gap geometry can make. The second section is a comparison of the computational model to experimental data. The results of the studies show that adding a simple fillet or chamfer to the downstream step of the gap can reduce the maximum heat flux by over 90%. These results can be used to reduce the heat transfer caused by flow impingement in the gaps of hypersonic vehicles with a simple modification of the geometry and is consistent with the findings of other work in gap heating.</p>

The study of the flow of a fan blade was conducted to improve tonal fan noise reduction by optimizing an existing flow control configuration. The current configuration consisted of a trailing edge Slot with a flow control area of 0.045 in² per inch span with an exit angle of -3.3° with respect to the blade exit angle. Two other flow control configurations containing discrete jets were investigated. For the first configuration, the trailing edge jets (TEJ), the fan blade was modified with discrete jets spaced 0.3 inches apart with a flow control area of 0.01 in² per inch span positioned on the trailing edge aimed at -3.3° with respect to the blade exit angle. Similarly, discrete jets were also placed on the suction surface at 95.5% chord aimed at 15° with respect to the local blade surface. This configuration is referred to as the suction surface jet (SSJ). The discrete jets for both configurations were designed to be choked while injecting a mass flow rate of 1.00% of the fan through-flow. Computational Fluid Dynamics (CFD) was used to model new configurations and study subsequent changes in total pressure deficit using a blade design inlet Mach number of 0.73, Reynolds number based on chord length of 1.67 à 106, and design incidence angle of 0°. Experimental testing was later conducted in a 2D cascade tunnel. The TEJ and SSJ were tested at design blowing of 1.00% and at off-design conditions of 0.50%, 0.75%, and 1.25% fan through-flow. Results between the different flow control configurations were compared using a blowing coefficient. CFD showed the TEJ and SSJ offered aerodynamic improvement over the Slot configuration. Testing showed the SSJ outperformed the TEJ, as validated in CFD, producing wider and shallower wakes. SSJ area-averaged pressure losses were 25% less than TEJ at design. Noise predictions based on CFD findings showed that both TEJ and SSJ provided additional tonal sound power level attenuation over the Slot configuration at similar blowing coefficients, with the SSJ providing the most attenuation. Noise prediction based on experimental results concurred that the SSJ provided more total attenuation than the TEJ. Experimental results showed that the SSJ performed better aerodynamically and, based on analytical prediction, provided 2 dB more total attenuation than the TEJ. / Master of Science

<p dir="ltr">This thesis explores novel methods of steel reheating, simulating hydrogen as a cleaner fuel in the combustion furnace and magnetic induction heating as a viable alternative, by utilizing advanced numerical simulations, including Computational Fluid Dynamics (CFD) and Finite Element Analysis (FEA), to assess their performance and feasibility.</p><p dir="ltr">Hydrogen, known for its potential to significantly reduce carbon dioxide emissions, is examined as a substitute for natural gas. Simulations revealed that hydrogen combustion results in higher flame temperatures and heat fluxes. While the CFD model achieved a high level of accuracy, with a maximum temperature error of 3% and an average deviation of 7% from real-world data, hydrogen fuel caused an increase in heat flux by up to 12% and higher slab surface temperatures. These changes led to steeper thermal gradients and increased stress, with peak stress levels reaching 90% of material limit. This simulation approach provides valuable data on the performance of different furnace fuels, helping to identify optimal fuel blends and configurations that minimize the risk of material failure while enhancing furnace efficiency.</p><p dir="ltr">The impact of scale formation on steel surfaces during reheating was also investigated. A mathematical model based on linear-parabolic equations was integrated into CFD simulations to predict scale growth. This model was validated against experimental data, showing an average error of 6%. The presence of scale led to a reduction in core temperature by up to 31 K and a 7.6% decrease in heat flux, which negatively affected heating efficiency. Scale formation also caused a significant drop in thermal conductivity, impacting heat transfer and slab uniformity. Pre-heating zone contributed minimally to overall scale formation despite its extended duration whereas a majority of scale growth was observed in the heating zone. Applications of this model include improving reheat furnace model efficiency and optimizing furnace operation to minimize scale.</p><p dir="ltr">Magnetic induction heating was also explored as an alternative to combustion-based reheating, assessing its potential benefits and challenges. The simulation results, validated with an average error of approximately 7% compared to literature data. showed uniform temperature distribution, and reduced stress levels with optimal power settings around 80 kW. A 3D transient simulation modeled an adaptive power cycle to minimize thermal stress highlighting the effectiveness of adaptive soaking strategies over continuous soaking in managing thermal stress, improving heating efficiency and material integrity.</p>

Doctor of Philosophy / Department of Mechanical and Nuclear Engineering / Zhongquan Zheng / A nano-material filter is an efficient device for improving indoor environmental quality (e.g. smoke reduction, air purification in buildings). Studying the effectiveness of nano-materials used in the device by computer simulation is challenging because very different size scales are involved. Therefore, numerical methods have to be developed to accommodate varying magnitudes of scales. In the current study, the simulation has been divided into three scales: macro-, micro- and nano-scale. The numerical schemes at each scale are targeted at a particular scale; however, the relationship of the general transport phenomena, physical mechanisms and properties among different scales are uniquely linked at the same time. The objective of the macro-scale simulation was to design and study a gas filter constructed with nano-material pellets. The filter was considered a packed-bed tube filled with manufactured nano-material pellets. Commercial computational fluid dynamics (CFD) packages were used along with the embedded programming macros. In the filtration process, we focused on the flow and species transport phenomena through the porous substrate. The mathematical/numerical models were built and tested based on the physical models used in the experimental setups for different materials that were tested. The results from the numerical models were validated and compared well to experimental data obtained from the pressure drop measurements and the adsorption (breakthrough) tests. In the micro-scale simulation, a modified immersed-boundary method (IBM) with the Zwikker-Kosten (ZK) porous model and the high-order schemes was validated and applied to simulate a representative porous unit that represented a periodic array of solid/porous cylinders. In the periodic unit, the solid cylinder case was used to validate the high-order schemes by comparing it to the results obtained from the commercial CFD software. The relationship between the pressure gradient and the porosity (Blake-Kozeny equation) was determined from this level and fed back to the macro-scale simulation, which provided a link between the two scales. In the porous cylinder case, both flow field and species transport were investigated with a porous model similar to the one used in the macro-scale. The species concentration change was calculated and found to be nonlinearly related to the adsorption coefficient. In the nano-scale simulation, a molecular dynamics (MD) simulation and a coupled molecular-continuum scheme were applied to solve the momentum and the mass transport problems at the molecular level at which the traditional continuum theory is no longer applicable. Both schemes were verified from the surface slip behavior study compared to the literature. The scale and shear effects in the Coutte flow were investigated, showing that in the micro-scale and macro-scale, the slip behavior could be neglected since it was only important in much smaller scales. The same hybrid scheme was then applied to a diffusion model with nano-pores constructed in the solid substrate. The adsorptions between various gases and the carbon substrate were simulated. The mass fluxes cross the fluid/solid interfaces were counted and both self-diffusivity and transport diffusivity were estimated and compared to their respective values found in the literature. The transport properties are closely related to the species transport (Fick’s law) in the macroscopic simulations. Linear concentration profiles in the channel were obtained based on those transport properties for various gases going through different sizes of nano-pores, which, as a connection to the continuum model, were to be used as boundary conditions in the continuum simulation.

Aujourd'hui divers systèmes de propulsion marins existent dont les systèmes de propulsion à jet. De nombreuses études ont été réalisées au cours des vingt dernières années et plus particulièrement par le biais de la simulation numérique ( Computational Fluid Dynamics , ( CFD )). Un modèle numérique simulant l'écoulement au travers un système de propulsion avec une pompe en rotation est développé dans ce projet. Il est également validé avec des résultats expérimentaux obtenus par d'autres chercheurs. Pour se [i.e. ce] faire, différentes étapes sont réalisées. En premier lieu, des paramètres tels que le maillage, le modèle de turbulence ou la modélisation de la rotation du rotor sont validés numériquement sur deux géométries. Le premier cas, bien documenté, correspond à un écoulement au travers une conduite en forme de"S" de section divergente et validé expérimentalement. Le second cas est une pompe où le rotor est en mouvement dans l'écoulement. Un modèle à multiple systèmes de référence ( Multiple Reference Frame,MFR ) est utilisé pour simuler la rotation de la pompe. Les paramètres numériques sont alors fixés pour la suite de l'étude. Ensuite un inodéle numérique d'un système de propulsion à jet d'une motomarine est développé dans un volume de contrôle réduit. Il prend en compte l'ensemble de la géométrie de la propulsion : la pompe en rotation, le venturi et une partie de la coque réelle du véhicule. Les conditions aux frontières sous la coque sont imposées grâce aux données d'une simulation complète du véhicule entier. Ce modèle est validé expérimentalement à deux vitesses (25 mph et 69 mph). Le comportement de l'écoulement est ensuite analysé. Finalement diverses variations géométriques sont effectuées telles que la supression [i.e. suppression] d'appendices dans la conduite ou le déplacement latéral de la lèvre de l'entrée d'eau à divers IVR, Inlet Velocity Ratio (rapport entre la vitesse du véhicule et celle de la pompe). Une augmentation des performances du système est observée à un certain IVR pour une supression [i.e. suppression] d'appendice donnée. La présence d'une plaque permettant le redressement de l'écoulement au niveau de la grille est néfaste à la poussée de même que la présence de l'arbre ou d'ailettes situées à l'entrée de la pompe. Le déplacement de la lèvre a pour objectif de déterminer la position optimale qui permet d'obtenir la meilleure augmentation de performance pour un IVR .