A presença de compostos orgânicos no solo proporciona o desenvolvimento de uma fase adicional nos poros, imiscível com a fase aquosa e gasosa pré-existentes. A característica de imiscibilidade destes compostos, denominados de fase livre não aquosa (“nonaqueos phase liquid” – NAPL), torna a avaliação de um solo contaminado por tais substâncias bastante complexa. Dentre as substâncias pertencentes a este grupo NAPL, destacam-se os produtos derivados do petróleo. As propriedades mais importantes na avaliação de um solo contaminado por NAPL são suas relações constitutivas. Uma destas propriedades relaciona a sucção dos fluídos com suas respectivas saturações, enquanto que a outra relaciona a condutividade relativa dos fluidos com a saturação de cada um destes fluidos. A partir da curva de retenção de um solo contaminado, é possível quantificar de forma indireta a relação condutividade relativa-saturação. Todavia, a obtenção desta curva de retenção para um solo contaminado por NAPL é bastante complexa e está em desenvolvimento, exigindo cuidados na sua determinação. Para isso, assume-se que no interior de um poro contendo água, óleo e ar, o contaminante estará preferencialmente localizado entre a água e o ar. Isto é determinante ao admitir-se que a saturação da água Sw é uma função da sucção mátrica desenvolvida entre a água e o óleo e a saturação total dos líquidos St, (soma da saturação da água e do óleo) é dependente da sucção mátrica gerada entre o óleo e o ar. Esta hipótese possibilita a obtenção das curvas de retenção para sistemas trifásicos ar-NAPL-água a partir dos sistemas bifásicos NAPL-água e ar-NAPL. Assim sendo, o objetivo deste trabalho é desenvolver, a partir de um oedômetro com controle de sucção, um equipamento, denominado de célula multifásica, que permita o controle independente das pressões do ar, do NAPL e da água. Especificamente, este equipamento permitirá a determinação das curvas de retenção para sistemas bi e trifásicos. Para atingir esta meta, foi necessário transformar uma pedra porosa cerâmica comum em uma pedra porosa hidrofóbica (repelente à água), através de um tratamento químico da pedra porosa convencional em uma solução contendo uma substância comercialmente conhecida como Glassclad 18. Este novo equipamento permite o monitoramento da variação do volume de água e de óleo no interior de uma amostra de solo para diversas pressões capilares. Ensaios preliminares foram realizados em uma mistura composta de 70% de um solo arenoso e 30% de um solo argilo-arenoso, sendo o primeiro coletado no estado do Rio Grande do Sul enquanto que o segundo em Santa Catarina. Ambos os solos possuem propriedades físicas bem conhecidas. / The presence of organic compounds in the soil provides the development of an additional phase in the pore space, immiscible with the preexisting water and gaseous phases. This characteristic of immiscibility of such substances, called nonaqueos phase liquids – NAPL’s, turns the evaluation of a soil contaminated by these substances sufficiently complex. Among the substances pertaining to this group, called NAPL, the petroleum products are distinguished. The most important properties in the evaluation of a soil contaminated by NAPL are their constitutive relationships. One of these properties relates the pressure of the fluids with their respective saturations, while the other one relates the relative permeability of the fluids with the saturation of each one of these fluids. From the pressure-saturation constitutive relationship of a contaminated soil, it is possible to quantify indirectly the relative permeability-saturation relationship. However, the evaluation of the pressure-saturation relationship for a soil contaminated by NAPL is sufficiently complex and is in development, demanding considerable care in its determination. For this, it is assumed that, in the interior of a pore contaning water, oil and air, the organic liquid will be preferentially located between the water and air. That is imperative when assuming that water saturation of Sw is a function of the suction developed between the oil-water interface and the total liquid saturation St (sum of water saturation and oil saturation) is dependent of the suction generated between the airoil interfaces. This hypothesis makes possible the evaluation of the pressure-saturation relationship for the three-phase system air-NAPL-water from the two-phase systems NAPLwater and air-oil. Thus, the objective of this work is to develop, from the edometer with pressure control, another equipment, called the multiphase cell, that allows the independent control of the air, NAPL and water pressures. Specifically, this equipment will allow the determination of the pressure-saturation relationship for bi- and three-phase systems. To reach this goal, it was necessary to transform a standard ceramic porous stone in a hidrophobic porous stone (repellent to water), through a chemical treatment of the standard porous stone in a solution containing a substance commercially known as Glassclad 18. This new equipment allows the control of the volume changes of water and oil within a soil sample for different suction. Preliminary tests have been carried out with a mixture composed by 70% of sand and 30% of silt; the first material were collected in the state of the Rio Grande do Sul (Brazil) and the second in the state of Santa Catarina. Both materials have their physical properties well known.

A presença de compostos orgânicos no solo proporciona o desenvolvimento de uma fase adicional nos poros, imiscível com a fase aquosa e gasosa pré-existentes. A característica de imiscibilidade destes compostos, denominados de fase livre não aquosa (“nonaqueos phase liquid” – NAPL), torna a avaliação de um solo contaminado por tais substâncias bastante complexa. Dentre as substâncias pertencentes a este grupo NAPL, destacam-se os produtos derivados do petróleo. As propriedades mais importantes na avaliação de um solo contaminado por NAPL são suas relações constitutivas. Uma destas propriedades relaciona a sucção dos fluídos com suas respectivas saturações, enquanto que a outra relaciona a condutividade relativa dos fluidos com a saturação de cada um destes fluidos. A partir da curva de retenção de um solo contaminado, é possível quantificar de forma indireta a relação condutividade relativa-saturação. Todavia, a obtenção desta curva de retenção para um solo contaminado por NAPL é bastante complexa e está em desenvolvimento, exigindo cuidados na sua determinação. Para isso, assume-se que no interior de um poro contendo água, óleo e ar, o contaminante estará preferencialmente localizado entre a água e o ar. Isto é determinante ao admitir-se que a saturação da água Sw é uma função da sucção mátrica desenvolvida entre a água e o óleo e a saturação total dos líquidos St, (soma da saturação da água e do óleo) é dependente da sucção mátrica gerada entre o óleo e o ar. Esta hipótese possibilita a obtenção das curvas de retenção para sistemas trifásicos ar-NAPL-água a partir dos sistemas bifásicos NAPL-água e ar-NAPL. Assim sendo, o objetivo deste trabalho é desenvolver, a partir de um oedômetro com controle de sucção, um equipamento, denominado de célula multifásica, que permita o controle independente das pressões do ar, do NAPL e da água. Especificamente, este equipamento permitirá a determinação das curvas de retenção para sistemas bi e trifásicos. Para atingir esta meta, foi necessário transformar uma pedra porosa cerâmica comum em uma pedra porosa hidrofóbica (repelente à água), através de um tratamento químico da pedra porosa convencional em uma solução contendo uma substância comercialmente conhecida como Glassclad 18. Este novo equipamento permite o monitoramento da variação do volume de água e de óleo no interior de uma amostra de solo para diversas pressões capilares. Ensaios preliminares foram realizados em uma mistura composta de 70% de um solo arenoso e 30% de um solo argilo-arenoso, sendo o primeiro coletado no estado do Rio Grande do Sul enquanto que o segundo em Santa Catarina. Ambos os solos possuem propriedades físicas bem conhecidas. / The presence of organic compounds in the soil provides the development of an additional phase in the pore space, immiscible with the preexisting water and gaseous phases. This characteristic of immiscibility of such substances, called nonaqueos phase liquids – NAPL’s, turns the evaluation of a soil contaminated by these substances sufficiently complex. Among the substances pertaining to this group, called NAPL, the petroleum products are distinguished. The most important properties in the evaluation of a soil contaminated by NAPL are their constitutive relationships. One of these properties relates the pressure of the fluids with their respective saturations, while the other one relates the relative permeability of the fluids with the saturation of each one of these fluids. From the pressure-saturation constitutive relationship of a contaminated soil, it is possible to quantify indirectly the relative permeability-saturation relationship. However, the evaluation of the pressure-saturation relationship for a soil contaminated by NAPL is sufficiently complex and is in development, demanding considerable care in its determination. For this, it is assumed that, in the interior of a pore contaning water, oil and air, the organic liquid will be preferentially located between the water and air. That is imperative when assuming that water saturation of Sw is a function of the suction developed between the oil-water interface and the total liquid saturation St (sum of water saturation and oil saturation) is dependent of the suction generated between the airoil interfaces. This hypothesis makes possible the evaluation of the pressure-saturation relationship for the three-phase system air-NAPL-water from the two-phase systems NAPLwater and air-oil. Thus, the objective of this work is to develop, from the edometer with pressure control, another equipment, called the multiphase cell, that allows the independent control of the air, NAPL and water pressures. Specifically, this equipment will allow the determination of the pressure-saturation relationship for bi- and three-phase systems. To reach this goal, it was necessary to transform a standard ceramic porous stone in a hidrophobic porous stone (repellent to water), through a chemical treatment of the standard porous stone in a solution containing a substance commercially known as Glassclad 18. This new equipment allows the control of the volume changes of water and oil within a soil sample for different suction. Preliminary tests have been carried out with a mixture composed by 70% of sand and 30% of silt; the first material were collected in the state of the Rio Grande do Sul (Brazil) and the second in the state of Santa Catarina. Both materials have their physical properties well known.

Le rôle des nuages sur la composition chimique atmosphérique est encore mal connu. Les composés chimiques présents sous forme de particules et de gaz sont transformés efficacement lors de leur passage dans le nuage par des processus photochimiques et microbiologiques. Les objectifs de ce travail de thèse étaient d’analyser l’efficacité de ces processus dans la transformation des composés chimiques à l’aide du modèle multiphasique de chimie du nuage CLEPS (CLoud Explicit Physico-chemical Scheme). Le premier travail a consisté à étudier la capacité oxydante du nuage sur la base de comparaisons entre des données expérimentales et simulées de vitesses de production du radical HO• pour des eaux nuageuses prélevées au puy de Dôme. Ces comparaisons ont montré que la photoréactivité du fer comme source de ce radical est surestimée par le modèle. Cela provient de la complexation du fer dans les échantillons naturels qui n'est pas considérée dans le modèle. Le modèle a également montré que la photolyse du peroxyde d’hydrogène représente la voie majoritaire dans la formation des radicaux HO•. Dans un second temps, les hypothèses formulées lors de la construction du nouveau mécanisme chimique en phase aqueuse CLEPS ont été validées par le biais d’expériences d’irradiation sur un composé cible, l’acide tartronique, réalisées en laboratoire et simulées avec le modèle. Le travail expérimental a confirmé la formation de l’acide glyoxylique prédite par le mécanisme en phase aqueuse du modèle. Ce dernier reproduit l’évolution temporelle des concentrations en acide tartronique et de son premier produit d’oxydation, l’acide glyoxylique mais sous-estime la formation de l’acide formique qui est un des produits finaux de l’oxydation de l’acide tartronique. Enfin, le modèle a été développé pour prendre en compte des vitesses de biodégradation mesurées en laboratoire. Le modèle permet de comparer l’efficacité des processus radicalaire et microbiologique dans la dégradation de quatre espèces chimiques en testant l’effet de paramètres environnementaux clés (température, flux actinique). L’activité biologique a été reconnue comme plus efficace dans le scénario estival que dans le cas hivernal et la contribution des microorganismes la nuit est dominante pour les deux scénarios. Un test de sensibilité a démontré que la contribution de l’activité microbienne dans la dégradation des composés chimiques est fortement accrue par rapport à la réactivité photochimique lorsque le fer est entièrement complexé. / The role of clouds on the atmospheric chemical composition is still poorly known. The chemical compounds under the form of particles and gases are efficiently transformed in the cloud by photochemical and microbiological processes. The thesis objectives were to analyze the efficiency of these processes in the transformations of the chemical compounds using the multiphase cloud chemistry model CLEPS (CLoud Explicit Physico-chemical schema). The first work consisted in studying the cloud oxidizing capacity based on comparisons between HO• radical production rates for irradiated cloud water sampled at the puy de Dôme station determined experimentally in the lab data and modeled with CLEPS. These comparisons showed that the model overestimates the contribution of the iron photolysis in the production of HO• radicals. This is due to the complexation of iron in cloud water samples that is not considered in the model due to the lack of data. The model also showed that the photolysis of hydrogen peroxide represents the major source of the HO• radicals in the aqueous phase. Secondly, the new protocol for cloud aqueous phase oxidation of organic compounds developed to build the CLEPS model was validated based on irradiation experiments of a target compound, the tartronic acid that were simulated by the model. The experimental work confirmed the formation of glyoxylic acid predicted by the aqueous phase mechanism in the model. It reproduces the temporal evolution of the concentrations of tartronic acid and of its first oxidation product, the glyoxylic acid but underestimates the formation of formic acid which is one of final products of the tartronic acid oxidation. Finally, the model was developed to take into account biodegradation rates measured in the laboratory. The model allows comparing the efficiency of the chemical and microbiological processes for the degradation of four chemical species and testing the effect of key environmental parameters (temperature, actinic flow). The biological activity was recognized as more effective in the summer case than in the winter case and the contribution of microorganisms the night is dominant for both scenarios. A sensitivity test demonstrated that the contribution of the microbial activity in the degradation of the chemical compounds is strongly increased in comparison with the photochemical reactivity when the iron is supposed to be totally complexed.

This thesis investigated the heat transfer of internally oil cooled rotors in permanent magnet electric machines which are, among other things, used in hybrid vehicles or zero emission vehicles. The magnets become sensitive and can be demagnetized at high working temperatures, hence the need of cooling. The scope of this work included CFD simulations in STAR-CCM+. Three different 3D multiphase models simulating the oil propagation in the rotor were performed. A Lagrangian multiphase model combined with a fluid film model was the most suitable model for simulating the spray of the oil and the film thickness along the inner rotor wall. It was noticed that periodic boundaries caused problems for the fluid film model, therefore a complete geometry was preferred over a truncated model. The 3D solutions provided thicker film thicknesses than the analytical solutions from the fluid film thickness theory. The maximum analytical thickness was of the same order of magnitude as the surface average film thickness provided by the multiphase models. This thickness was assumed to be constant when used as the base for the fluid region in the 2D one-phase models.The study showed that aluminum was the most suitable rotor material due to its high conductive capacity, which provided a more even distribution of the temperature in the solid and hence resulted in lower overall temperatures. The cooling power increased linearly with the volumetric flow rate, however the heat transfer coefficient decreased for the higher flow rates. A volumetric flow rate of 10dl/min was recommended. A 2D model was compared to a preliminary experiment and showed that these were not correlated. The conclusion was that more experiments and simulations are needed in order to confirm the validity of the 2D model.

As a scalable, high-efficiency, and simple converter topology, an interleaved, multiphase buck converter has been widely used to power microprocessors in information industry. As modern microprocessors continuously advance, the required current for high-performance microprocessors used in data center applications could be several hundreds of amperes with a current slew rate larger than 1000 A/μs. This poses great challenges for a high-efficiency, high-power-density voltage regulator design with a fast transient response. On the other hand, the design challenges of voltage regulators in mobile applications are also increasing due to the stringent requirement on the device thickness and the battery life. In a multiphase buck converter, discrete inductors are widely used as energy storage elements. However, this solution has a limited transient response with a large size of magnetic components. To overcome these issues, coupled inductor is proposed to realize a small steady-state current ripple, a fast transient response, and a small inductor size at the same time. Although lots of studies have been conducted in the topic of the coupled inductor, there are still several challenges unsolved in this area. These challenges are addressed through a comprehensive study in this dissertation. First, a comprehensive analysis of different coupled inductor structures is crucial to identify the benefits and limitations of each inductor structure and provide design guidance under different application requirements. Based on the coupling mechanism, different coupled inductor structures are categorized as a direct-coupled inductor (DCL), an indirect-coupled inductor (ICL) or a hybrid-coupled inductor (HCL) in this work. The performance of these three types of coupled inductors is analyzed in detail through the equivalent inductance analysis and the magnetic flux analysis. For the applications that require a small phase number, a DCL can achieve the smallest inductor size with a given inductance requirement. As the phase number increases, it is beneficial to use an ICL and an HCL due to their symmetrical, simple, and scalable inductor structures. As compared to an ICL, an HCL can achieve a smaller inductor size due to the flux-cancellation effect. The difference between a DCL, an ICL and an HCL are revealed quantitively with several design examples through this study. Second, the steady-state inductance (Lss) and the transient inductance (Ltr) are two key design parameters for coupled inductors. A large Lss and a small Ltr are preferred from the circuit performance point of view. However, there is a design conflict in an ICL and an HCL under the inductor size constraint, where reducing Ltr also results in a smaller Lss. A variable coupling coefficient concept is proposed to overcome this issue. With the same Lss, the proposed method can achieve a smaller Ltr during load transients as compared with the conventional method. This concept is realized by applying a nonlinear inductor in the additional winding loop with the current in this loop as the control source. Compared with the conventional structure, the proposed structure can achieve a great output voltage spike reduction and output capacitance reduction. Third, although an ICL and an HCL are promising candidates for multiphase coupled inductors, an extra inductor is required in the additional winding loop to adjust the coupling coefficient. This additional inductor occupies extra space. To shrink the total inductor size, several improved magnetic core structures are proposed to achieve the controllable coupling through the magnetic integration for an ICL and an HCL. Furthermore, the thickness of the core plate can be significantly reduced by the improved core structure for an HCL. Overall, it is demonstrated that the inductor footprint is greatly reduced by the proposed core structure, as compared with the conventional solution. Lastly, a novel PCB-embedded coupled inductor structure is proposed for a 20MHz integrated voltage regulator (IVR) for mobile applications. To achieve a small inductor footprint and a low profile, the inductor structure with a lateral flux pattern and direct coupling is adopted. Compared with the state-of-the-art solution, the proposed structure can adjust the coupling in a simple core structure by changing the inductor winding pattern. The proposed structure integrates multiple inductors into one magnetic core and is embedded into PCB with a total thickness of 0.54 mm. In contrast to prior arts, the proposed inductor structure features a large inductance density and quality factor with a much smaller DC resistance (DCR), thus is seen as a promising candidate for IVR applications. / Doctor of Philosophy / As modern microprocessors continuously advance in the information industry, the required current for high-performance microprocessors used in data center applications could be several hundreds of amperes with a current slew rate larger than 1000 A/μs. This poses great challenges for the power converter design. On the other hand, the design challenges of power converters in mobile applications are also increasing due to the stringent requirement on the device thickness and the battery life. As a scalable, high-efficiency, and simple converter topology, an interleaved, multiphase buck converter has been widely used to power these processors. In a multiphase buck converter, discrete inductors are widely used as energy storage elements. However, this solution has a limited transient response with a large size of magnetic components. To overcome these issues, coupled inductor is proposed to realize a small steady-state current ripple, a fast transient response, and a small inductor size at the same time. Although lots of studies have been conducted in the topic of the coupled inductor, there are still several challenges unsolved in this area. These challenges are addressed through a comprehensive study in this dissertation. First, a comprehensive analysis and comparison of different coupled inductor structures is crucial to identify the benefits and limitations of each inductor structure and provide design guidance under different application requirements. Based on the coupling mechanism, different coupled inductor structures are categorized as a direct-coupled inductor (DCL), an indirect-coupled inductor (ICL) or a hybrid-coupled inductor (HCL) in this work. The performance of these three types of coupled inductors is analyzed in detail through the equivalent inductance analysis and the magnetic flux analysis. The difference between a DCL, an ICL and an HCL are revealed quantitively with several design examples through this study. Second, the steady-state inductance (Lss) and the transient inductance (Ltr) are two key design parameters for coupled inductors. A large Lss and a small Ltr are preferred from the circuit performance point of view. However, there is a design conflict in an ICL and an HCL under the inductor size constraint, where reducing Ltr also results in a smaller Lss. A variable coupling coefficient concept is proposed to overcome this issue. This concept is realized by applying a nonlinear inductor in the conventional structure. Compared with the conventional structure, the proposed structure can achieve a great output voltage spike reduction and output capacitance reduction. Third, although an ICL and an HCL are promising candidates for multiphase coupled inductors, an extra inductor is required in the additional winding loop to adjust the coupling coefficient. This additional inductor occupies extra space. To shrink the total inductor size, several improved magnetic core structures are proposed to achieve the controllable coupling through the magnetic integration for an ICL and an HCL. Lastly, a novel PCB-embedded coupled inductor structure is proposed for a 20MHz integrated voltage regulator (IVR) for mobile applications. Compared with the state-of-the-art solution, the proposed structure can adjust the coupling in a simple core structure by changing the inductor winding pattern. In contrast to prior arts, the proposed inductor structure features a large inductance density and quality factor with a much smaller DC resistance (DCR), thus is seen as a promising candidate for IVR applications.

There are many options available when selecting a computational model for two-phase flows. It is important to understand all the features of the model selected, including when the model is appropriate and how using it may affect your results. This work examines how volume displacement effects in two-phase Eulerian-Lagrangian models manifest themselves. Some test cases are examined to determine what input these effects have on the flow, and if we can predict when they will become important. Bubble injection into a traveling vortex ring is studied in-depth, as it provides significant insight into the physics of these volume displacement effects. When a few bubbles are entrained into a traveling vortex ring, it has been shown that even at extremely low volume loadings, their presence can significantly affect the structure of the vortex core (Sridhar & Katz 1999). A typical Eulerian-Lagrangian point-particle model with two-way coupling for this dilute system, wherein the bubbles are assumed subgrid and momentum point-sources are used to model their effect on the flow, is shown to be unable to accurately capture the experimental trends of bubble settling location, bubble escape, and vortex distortion for a range of bubble parameters and vortex strengths. Accounting for fluid volume displacement due to bubble motion, using a model termed as volumetric coupling, experimental trends on vortex distortion and bubble settling location are well captured. The fluid displacement effects are studied by introducing the notion of a volume displacement force, the net force on the fluid due to volumetric coupling, which is found to be dominant even at the low volume loadings investigated here. A method of quantifying of these forces is derived and used to study the effects for a wide range of particle to fluid density ratios in Taylor-Green vortices. A simple modification to the standard point-particle Lagrangian approach is developed, wherein the interphase reaction source terms are consistently altered to account for the fluid displacement effects and reactions due to bubble accelerations. / Graduation date: 2013 / Access restricted to the OSU Community at author's request from Sept. 24, 2012 - Sept. 24, 2013

Computational Fluid Dynamics (CFD) has become one of the main instruments for the prediction of many commercial and research oriented fluid flow and heat transfer problems. While single phase flow analysis through CFD has gained grounds within the commercial industry, multiphase flow analysis is still the subject of further research and development. Heat Pipes and thermosyphones are no exception to this. However, the involvement of more than one fluid phase within these devices has made their analysis through CFD more challenging and computationally more demanding to perform. In this thesis, computational fluid dynamics is used as a modelling tool in order to predict the thermal hydraulic behaviour of multiphase environment within thermosyphones and heat pipes. Eulerian two-fluid method is used to solve the conservation equations for mass, momentum and energy, for each phase along with the inclusion of interfacial heat and mass transfer terms. Numerical predictions are obtained for the steady-state and transient operation of stationary thermosyphon, while rotating heat pipes operation is also simulated using axially and radially rotating heat pipe models. Apart from using the commercially available CFD code for the analysis of thermosyphones related simulation, numerical work is performed regarding the coupling of momentum equations based on Eulerian two-fluid modelling scheme. OPENFOAM open source code is used and modified to include the Partial Elimination Algorithm (PEA) for the coupling of interfacial exchange terms, including interfacial mass transfer term, in the momentum equations of both phases. Results obtained from above discussed studies provide good agreement with corresponding experimental and analytical observations.

Sand transport in hilly terrain geometry is different and complex to understand compared to horizontal pipeline, due to the influence of the geometry that greatly affect multiphase flow and sand behaviour at the dip. The overall aim of this research work is to use experimental method to investigate the effects of multiphase flow behaviour on sand transport in a dip configuration. Experimental work was carried out to understand the complex dynamic mechanisms that exist during sand multiphase flow using 2” inch dip test facility with different inclination angles of ±24° and ±12° configurations. In order determine the multiphase flow parameters and sand flow regimes, extensive data were collected and analysed from continuous local measurement of instantaneous liquid hold up and sand hold up using conductivity rings. Significant observations were made during this study from which several conclusions were made. In the air–water test, three slug behaviours were observed at the dip: complete stratified flow downhill with slug initiation at dip; stratified flow (with energetic ripple) downhill with slug initiation and slug growth upward dip; and aerated slug downhill and slug growth at the dip. These behaviours are different from published work on this subject with low angle of inclination. The data analysis revealed that the two types of slug initiation mechanisms (wave growth and wave coalescence) observed are geometry specifics. The slug translational velocities (at the dip and uphill section) were used as criterion to determine the flow condition for each slug initiation mechanism at the dip. Significant observations were made during this study from which several conclusions were made. In the air–water test, three slug behaviours were observed at the dip: complete stratified flow downhill with slug initiation at dip; stratified flow (with energetic ripple) downhill with slug initiation and slug growth upward dip; and aerated slug downhill and slug growth at the dip. These behaviours are different from published work on this subject with low angle of inclination. The data analysis revealed that the two types of slug initiation mechanisms (wave growth and wave coalescence) observed are geometry specifics. The slug translational velocities (at the dip and uphill section) were used as criterion to determine the flow condition for each slug initiation mechanism at the dip. Five sand-water flow regimes (full suspension, streak, saltation, sand dune, and sand bed) were established by physical observation and data analysis. It was also observed that sand streaks were denser towards the central line of pipe bottom in the downhill pipe than that in uphill pipe. At downhill pipe section, there were sand gathering toward the central line of the pipe bottom. The characteristics of sand transportation at the dip section were found slightly different from downhill and uphill pipe for higher sand concentrations. When dense streak occurred at the downhill, the sand particles become dispersed at the dip. The minimum transport conditions (MTC) were determined at different sand concentration. The sand minimum transport condition in the dip section was found to be slightly lower than those in the downhill and uphill section. The minimum transport condition for a single phase water flow for the 24˚ dip. test section was slightly higher (with difference of about 0.1m/s) than that of the 12˚ at the downward and upward of the dip section at low sand concentration. In addition, local sand measurements using conductivity time series results at the downhill and uphill section showed the influence of sand concentration and flow condition on sand flow patterns. The air-water-sand results showed that sand deposits occurred in uphill section after sand transport at the downhill and dip sections. This was due to different flow regimes exhibited at the different pipe sections. The stratified (wavy) flow was the dominant flow in downhill pipe; therefore sand was observed transporting within the liquid film as thin streak for most of test conditions. The slug initiation at the dip section was observed to be a major factor that influences the sand behaviour. Sand particles in the slug unit (at the dip and uphill of the pipe) were observed to be entrained in the slug body once slug is initiated, thereby changing the force vector generating turbulence flow at the front of slug body. Once the sand particles entered the film zone of the slug unit, they immediately stopped moving forward due to the film velocity significantly lower than the slug body coupled with gravity effect. . Sand particles were found to be falling back while travelling with the water film at some conditions, until they were picked up by the next slug body. The results of this work provide a better understanding to the study of multiphase flow for pipeline design and most especially sand behaviour at the dip. The sand dune regime is identified distinctively using conductivity ring technique which would assist in determining the operating conditions that allow sand dune formation. The knowledge of flow condition at full suspension of sand is an important parameter to determine the erosion rate over the life span of the pipeline. Also, the quantity of sand bed and flow condition of sand settling at the dip is useful information for production chemist in order to determine the effectiveness of corrosion inhibitor at the bottom of the pipe. In conclusion, sand transport characteristics and MTC were strongly dependent on the gas-liquid flow regime and pipe geometry; and cannot be generalised on the superficial liquid and gas velocities of the transport fluid.

Accurate predictions of macroscopic multiphase flow properties, such as relative permeability and capillary pressure, are necessary for making key decisions in reservoir engineering. These properties are usually measured experimentally, but pore-scale network modeling has become an efficient alternative for understanding fundamental flow behavior and prediction of macroscopic properties. In many cases network modeling gives excellent agreement with experiment by using models physically representative of real media. Void space within a rock sample can be extracted from high resolution images and converted to a topologically equivalent network of pores and throats. Multiphase fluid transport is then modeled by imposing mass conservation at each pore and implementing the Young-Laplace equation in pore throats; the resulting pressure field and phase distributions are used to extract macroscopic properties. Advancements continue to be made in making network modeling predictive, but one limitation is that artificial (e.g. constant pressure gradient) boundary conditions are usually assumed; they do not reflect the local saturations and pressure distributions that are affected by flow and transport in the surrounding media. In this work we demonstrate that flow behavior at the pore scale, and therefore macroscopic properties, is directly affected by the boundary conditions. Pore-scale drainage is modeled here by direct coupling to other pore-scale models so that the boundary conditions reflect flow behavior in the surrounding media. Saturation couples are used as the mathematical tool to ensure continuity of saturations between adjacent models. Network simulations obtained using the accurate, coupled boundary conditions are compared to traditional approach and the resulting macroscopic petrophysical properties are shown to be largely dependent upon the specified boundary conditions. The predictive ability of network simulations is improved using the novel network coupling scheme. Our results give important insight into upscaling as well as approaches for including pore-scale models directly into reservoir simulators. / text

A series of local and bench scale laboratory experiments and bench and field scale numerical simulations were conducted to develop a better understanding of the interrelationship between nonwetting phase (NWP) source zones and downgradient aqueous phase concentrations in saturated porous media contaminated by immiscible organic liquids. Specific emphasis was placed on the factors governing the rate of NWP source zone evolution and the factors governing the rate of mass transfer from the NWP to the aqueous phase. Hysteretic NWP relative permeability-saturation (krN-SW) relationships were measured at the local scale for six sands to examine the relationship between krN-SW functions and porous media type. Parameterization of the measured constitutive relationships revealed a strong correlation between mean grain diameter and the maximum value of NWP relative permeability. The measured krN-SW relationships, were validated through a bench scale experiment involving the infiltration, redistribution, and immobilisation of NWP in an initially water saturated heterogeneous porous medium. This match of simulation to experiment represents the first validation of a multiphase flow model for transient, fixed volume NWP releases. Multiphase flow simulations of the bench scale experiment were only able to reproduce the experimental observations, in both time and space, when the measured krN-SW relationships were employed. Two-dimensional field scale simulations of a fixed volume NWP release into a heterogeneous aquifer demonstrate the influence of spatially variable krN-S relationships correlated to porous media type. Both the volume of the NWP invaded porous media, and the length of time during which NWP is migrating, will be under predicted if variable (correlated) kr,N is not accounted for in the numerical model iv formulation. This under prediction is exacerbated as the mean intrinsic permeability of the release location decreases. A new, thermodynamically-based interfacial area (IFA) model was developed for use in the single-boundary layer expression of mass transfer as an alternative to existing empirical correlation expressions. The IFA model considers consistency and continuity of constitutive relationships, energy losses, effective specific interfacial area for mass transfer, and dissolution of residual NWP. A bench scale experiment involving the release and dissolution of a transient NWP source zone in heterogeneous porous media was conducted to evaluate the appropriateness of the developed IFA model when utilised to predict NWP dissolution rates. Comparison of measured downgradient dissolved phase concentrations and source zone NWP saturations in time and space with those from numerical simulations of the experiment reveal that the proposed IFA model is superior to both a local equilibrium assumption and existing empirical correlation expressions. This represents the first mass transfer model validated for the dissolution of a complex NWP source zone. Twodimensional simulations at the field scale of multiphase flow and dissolution suggest that employing existing mass transfer expressions instead of the IFA model lead to incorrect predictions of the life spans of NWP source zones, downgradient dissolved phase concentrations, and the rate of mass flux through a downgradient boundary. The practical implication of this research is that accurate numerical predictions of the evolution of a transient NWP source in porous media require consideration of krN-S relationships and NWP / aqueous phase IFA, as these factors dictate the rates of the key subsurface contaminant processes of migration and dissolution, respectively.