[pt] A injeção de emulsão em meio poroso como método de recuperação avançada pode se tornar realidade na operação de campos de petróleo devido à maior rigidez no descarte de água produzida e aos potenciais ganhos na produção de óleo. Para entender o comportamento macroscópico desta técnica de EOR, é necessário compreender o fenômeno microscópico. Com este objetivo, fez-se a modelagem numérica do escoamento de uma gota imersa em fase contínua escoando em capilar restrito utilizando o método de Lattice-Boltzmann. Este método foi escolhido devido à sua facilidade de ser aplicado em geometrias complexas de rocha e fluido e ao bom compromisso na representação dos fenômenos de microescala. Para entender a influência de cada parâmetro, foram realizadas diversas simulações em domínio tridimensional, alterando a velocidade do fluxo, a razão de viscosidades dos fluidos, a relação entre os diâmetros da gota e do tubo e a magnitude da tensão interfacial. Observou-se que a passagem da gota pela restrição causa uma redução na mobilidade do escoamento, representada por um aumento na perda de carga, pela conjunção dos efeitos viscosos e capilares. Obteve-se correlação negativa do fator redutor de mobilidade 𝑓 com a razão de viscosidade e com o tamanho da gota, assim como fora determinado numericamente por Roca-Reyes (2011) com o método level-set. Foi notada uma pequena sensibilidade de 𝑓 ao número de capilaridade, assim como estabelecido experimentalmente por Robles-Castillo (2011). Verificou-se a importância de se determinar o conjunto adequado de parâmetros do sistema para ter sucesso na implantação de injeção de emulsões. / [en] Emulsion injection in porous medium as an Enhanced Oil Recovery method can turn out to be reality in the operation of onshore and offshore fields, due to increasing rigidity in the disposal of produced water and also due to the potential additional oil production. In order to understand macroscopic behavior of this EOR method, it is necessary to understand the microscopic phenomenon. With this objective, it was performed the numerical modeling of the flow of a droplet immerse in continuous phase through a constricted capillary using the Lattice-Boltzmann method. This method was chosen due to its simplicity on being applied to complex rock geometries and multiphasic flow and due to its good commitment in representing microscopic phenomena. Focusing on understanding the influence of each parameter on flow behavior, several simulation studies were performed altering flow velocity, viscosity ratio, ratio between droplet s and pipe s diameter and interfacial tension. A reduction in mobility is observed as the droplet passes through the restriction caused by the conjunction of viscous and capillary effects. A negative correlation of mobility reduction factor 𝑓 in relation to the viscosity ratio and to droplet size was noticed, as it had been observed before by Roca-Reyes (2011) in a numerical implementation of level-set method. Weak correlation between 𝑓 and capillary number was determined, as in previous experimental essay performed by Robles-Castillo (2011). In this study, it was verified the importance of determining the appropriate set of system parameters, in order to achieve success in the implementation of emulsion injection.

Concerns about the use of certain chemical species within the aerospace field are growing in recent years. A European regulation, REACh, now makes the use of hydrazine uncertain in – among others- attitude control thrusters. Green monopropellants, which are alternatives for this species already exist, but they all require a catalyst to react. Catalysts constitute the limiting factor for the lifespan of satellites because of the number of thermal cycles they endure. A joint project between ONERA, the French aerospace research center and CNES, the French space agency, was born to develop a high-performance green monopropellant thruster operating without any catalyst. Sizing the thruster and particularly its combustion chamber is not an easy task because of the explosive properties and the lack of knowledge regarding the monopropellant reaction process. The thesis aims at simulating the flow in a combustion chamber using CNES05, a new promising green monopropellant. This monopropellant has a very low vapor pressure and is an energetic liquid. As such, its reaction above a certain temperature -which is called decompositionis not well understood and must be observed closely. For this matter, a test bench was created, and it paved the way for the development of a specific model of decomposition. Indeed, even if the CNES05 decomposition cannot be modeled with the classical theory of isolated droplets, the setup showed us the order of magnitude of the reaction kinetics and the presence of a break up phenomenon. Using this model, the simulations of the flow inside the combustion chamber give us the heat flux profile through its walls, a sizing parameter for the thruster. Large recirculation zones are observed and the influence of the angle of injection seems to be the major injection parameter of influence. The sensitivity of the parameters used in the model is also studied.

Gas-liquid two-phase flows occur in many industrial applications and apparatuses. The design and optimization of such apparatuses and processes requires the numerical simulation of two-phase flows. However, two-phase flow simulations are still a challenging task, especially for industrial scales. Here, the simulation of large flow domains and high Reynolds number flows require a reduction of the resolved time-scales and length-scales by a high level of modeling to decrease the computational effort. Therefore, physics-based models are needed to depict the complex transport processes between the phases. Thus, two-phase flows are the object of ongoing research. Up to now, the majority of closure models for turbulence, interfacial forces or bubble breakup and coalescence were validated against experimental data derived from experiments in simple flow channel geometries like straight pipes. Their application for the simulation of two-phase flows with three-dimensional flow structures like e.g. recirculating areas, flow separation or strong velocity gradients requires constant experimental validation and further development. Hence, improved experimental methods are required for investigations of gas-liquid two-phase flows to provide reliable data for further development and validation of numerical flow simulation models. Therefore, experiments were performed in a constricted pipe under bubbly two-phase flow conditions. Three-dimensional flow structures were created by two types of flow constrictions for a variety of gas and liquid superficial velocities up to jg = 0.1400 m⋅s-1 and jl = 1.6110 m⋅s-1. The flow fields upstream and downstream of the flow constrictions were studied using ultrafast X-ray tomography and hot-film anemometry to obtain cross-sectional phase distribution, bubble characteristics and liquid velocity. The analysis of the ultrafast X-ray tomography image data was significantly improved by development of a histogram-based gas holdup calculation. Furthermore, the spatial dependence of the axial image plane distance was studied to improve the determination of axial bubble velocities and, thus, bubble sizes. The experimental method was advanced by simultaneous application of ultrafast X-ray tomography and hot-film anemometry. Eventually, the experimental data was compared to state-of-the-art Euler/Euler two-fluid simulations. The simulations were performed in the framework of a parallel doctoral thesis in the Experimental Thermal Fluid Dynamics department at the Helmholtz-Zentrum Dresden – Rossendorf by Ms. Sibel Tas-Koehler following the baseline approach. The results were compared in terms of the phase distribution, bubble sizes and gas velocity for two operating conditions using the homogeneous multiple size group model. / Zweiphasenströmungen aus Gasen und Flüssigkeiten treten in vielen industriellen Anwendungen und Apparaten auf. Um einen sicheren, zuverlässigen und optimalen Betrieb einzelner Komponenten und gesamter Anlagen zu gewährleisten, sind die Strömungen Gegenstand zahlreicher Untersuchungen. Numerische Simulationen sind ein unverzichtbares Instrument, um Prozesse unter diesen Aspekten zu bewerten. Die Simulation von Zweiphasenströmungen, insbesondere im industriellen Maßstab, ist jedoch nach wie vor eine anspruchsvolle Aufgabe. Um den Rechenaufwand zu verringern und die Simulation von großen Strömungsgebieten und Strömungen mit hohen Reynoldszahlen zu ermöglichen, ist ein hohes Maß an Modellierung notwendig. Gleichzeitig wurden die meisten Schließungsmodelle zur Beschreibung von Turbulenz, Grenzflächenkräften oder Blasenzerfall und -koaleszenz für einfache Geometrien wie beispielsweise gerade Rohre entwickelt. Die Anwendung dieser Modelle für die Simulation von Zweiphasenströmungen mit dreidimensionalen Strömungsstrukturen, wie z.B. Rezirkulationsgebieten, Strömungsablösungen oder starken Geschwindigkeitsgradienten, erfordert eine ständige experimentelle Validierung und Weiterentwicklung. Dies wiederum erfordert eine immer höhere Auflösung der eingesetzten Messsysteme und steigende Qualität der experimentellen Daten. Um verlässliche Daten für die Weiterentwicklung und Validierung von Modellen für die numerische Strömungssimulation zu erhalten sind daher verbesserte experimentelle Methoden zur Untersuchung von Gas-Flüssig-Strömungen erforderlich. Aus diesem Grund wurden Experimente an einer Blasenströmung in einem Rohr mit einer Strömungsverengung durchgeführt. Zwei Arten von Verengungen wurden genutzt, um dreidimensionale Strömungsstrukturen für eine Vielzahl von Betriebsbedingungen zu erzeugen. Diese sind durch Gas- und Flüssigkeitsleerrohrgeschwindigkeiten bis zu jg = 0.1400 m⋅s-1 und jl = 1.6110 m⋅s-1 definiert. Um die Phasenverteilung im Querschnitt der Strömung, Blaseneigenschaften und die Flüssigphasengeschwindigkeit stromauf- und -abwärts der Verengung zu ermittelt, wurde die Strömung mit Hilfe der ultraschnellen Röntgentomographie und Heißfilm-Anemometrie untersucht. Die Datenanalyse für die Bilddaten der ultraschnellen Röntgentomographie wurde durch die Entwicklung einer Histogramm-basierten Gasgehaltsberechnung erheblich verbessert. Um die Bestimmung der axialen Blasengeschwindigkeiten und damit der Blasengrößen zu verbessern, wurde außerdem die räumliche Abhängigkeit des axialen Bildebenenabstands untersucht. Die experimentellen Methoden wurden durch die gleichzeitige Anwendung von ultraschneller Röntgentomographie und Heißfilm-Anemometrie weiterentwickelt. Die experimentellen Daten wurden mit dem Stand der Technik von Euler/Euler-Zweiphasen-Simulationen verglichen. Die Simulationen wurden im Rahmen eines parallelen Promotionsvorhabens in der Abteilung Experimentelle Thermofluiddynamik am Helmholtz-Zentrum Dresden – Rossendorf von Frau Sibel Tas-Köhler durchgeführt und folgten der Baseline-Modell Strategie. Die Ergebnisse wurden unter Verwendung des homogenen Modells mehrerer Größenklassen bezüglich der Phasenverteilung, der Blasengrößen und der Gasgeschwindigkeit für zwei Betriebsbedingungen verglichen.

Gas-liquid two-phase flows are encountered in different industrial applications such as, chemical reactors, wastewater treatment, oil and gas exploration and nuclear reactors. In nuclear reactors, boiling two-phase flows occur under both normal and accident conditions. For the design and safety operation of nuclear reactors, Computational Fluid Dynamics (CFD) based on the Euler−Euler framework has become a popular tool. However, accurate CFD prediction for a fuel assembly geometry is still a challenge. The reason is that the accuracy of two-phase flow simulations is highly dependent on adequate modelling of phase interactions including interfacial forces (i.e. drag, lift, wall lubrication, turbulent dispersion and virtual mass), bubble-induced turbulence (BIT) and bubble breakup/coalescence. Through the Euler−Euler framework, modelling of these phase interactions is provided by different approaches. These approaches include closure equations, most of which have been determined empirically. These closures are important for the accurate prediction of mean flow profiles, including void fraction and phase velocity distributions. A variety of closure models has been proposed by different researchers. However, it is difficult to differentiate them and make an appropriate choice for a particular problem without knowing their predictive properties in detail. While an extensive number of models have been developed and have meanwhile been well validated for simple pipe and column geometries, there is yet limited analysis and qualification for more complex three-dimensional flow domains. One reason for this is the lack of suitable experimental validation data. In addition, it is important to mention that most of the available models were generally obtained considering laminar or low turbulence conditions. Therefore, it is necessary to further investigate the modelling capabilities for two-phase flows with flow complexity/high turbulence as they occur in nuclear reactors. For this purpose, additional validations are required in the CFD modelling of two-phase flows. However, studies on the capabilities of two-phase flow models directly for rod bundles are very complicated and time-consuming. Hence, a capability analysis of the models for the three main phenomena, i.e. breakup/coalescence, drag and turbulence, was first carried out for the case of a semi-obstructed pipe under adiabatic flow conditions. The results were validated using the experimental data obtained by Neumann-Kipping (2022) on the void fraction, mean bubble diameter, bubble size distribution, liquid velocity and gas velocity for two different turbulence conditions. Subsequently, experiments were conducted in a 3 x 3 rod bundle with a spacer and vanes using X-ray computed tomography (CT), which provides high quality void data without disturbing the flow. The effects of different mass and heat fluxes on the void fraction and its distribution downstream of the spacer were analyzed. In addition, the effects of different vane angles on the distribution of the void fraction were discussed. Furthermore, an experimental database was obtained in a rod bundle with a spacer under different flow conditions to validate the numerical modelling. Finally, the improved CFD model obtained from the semi-obstructed pipe geometry was applied to the 3 x 3 rod bundle geometry under two different turbulence conditions. The numerical results were compared with the X-ray CT data on the void fraction. / Gas-Flüssig-Zweiphasenströmungen kommen in verschiedenen industriellen Anwendungen wie Blasensäulen, Rührkesseln und Kernreaktoren vor. In Kernreaktoren treten siedende Zweiphasenströmungen sowohl unter Normal- als auch unter Störfallbedingungen auf. Für die Auslegung und den sicheren Betrieb von Kernreaktoren ist die numerische Strömungsmechanik (engl. Computational Fluid Dynamics, CFD) auf der Grundlage des Euler−Euler-Konzepts zu einem wichtigen Instrument geworden. Eine genaue CFD-Vorhersage für eine Brennelementgeometrie ist jedoch nach wie vor eine Herausforderung. Der Grund dafür ist, dass die Genauigkeit von Zweiphasenströmungssimulationen in hohem Maße von einer genauen Modellierung der Phasenwechselwirkungen abhängt, einschließlich der Grenzflächenkräfte (d. h. Widerstand, Lift, Wand, turbulente Dispersion und virtuelle Masse), der blaseninduzierten Turbulenz (BIT) und des Aufbrechens/Koaleszierens von Blasen. Durch den Euler−Euler-Rahmen wird die Modellierung dieser Phasenwechselwirkungen durch verschiedene Ansätze ermöglicht. Zu diesen Ansätzen gehören Schließungsgleichungen, von denen die meisten empirisch ermittelt wurden. Diese Schließungsgleichungen sind wichtig für die genaue Vorhersage von mittleren Strömungsprofilen, einschließlich Gasgehalt und Phasengeschwindigkeitsverteilungen. Es gibt eine Vielzahl von Schließungsmodellen, die von verschiedenen Forschern innerhalb ihrer experimentellen Bereiche vorgeschlagen wurden. Es ist jedoch schwierig, sie zu unterscheiden und eine geeignete Wahl für ein bestimmtes Problem zu treffen, ohne ihre Vorhersageeigenschaften im Detail zu kennen. Während für einfache Rohr- und Säulengeometrien eine große Anzahl von Modellen entwickelt und inzwischen gut validiert wurde, gibt es für komplexere dreidimensionale Strömungsgebiete noch wenig Analyse und Qualifizierung. Ein Grund dafür ist der Mangel an geeigneten experimentellen Validierungsdaten. Darüber hinaus ist es wichtig zu erwähnen, dass die meisten der verfügbaren Modelle im Allgemeinen unter laminaren oder geringen Turbulenzbedingungen erstellt wurden. Daher ist es notwendig, die Modellierungsmöglichkeiten für Zweiphasenströmungen mit komplexer Strömung/hoher Turbulenz, wie sie in Kernreaktoren auftreten, weiter zu untersuchen. Zu diesem Zweck sind zusätzliche Validierungen bei der CFD-Modellierung von Zweiphasenströmungen erforderlich. Untersuchungen zur Leistungsfähigkeit von Zweiphasenströmungsmodellen direkt für Stabbündel sind jedoch sehr kompliziert und zeitaufwändig. Daher wurde zunächst eine Fähigkeitsanalyse der Modelle für die drei Hauptphänomene, d. h. Aufbrechen/Koaleszenz, Widerstand und Turbulenz, für den Fall eines halbgeschlossenen Rohrs unter adiabatischen Strömungsbedingungen durchgeführt. Die Ergebnisse wurden anhand der von Neumann-Kipping (2022) gewonnenen experimentellen Daten über den Gasgehalt, den mittleren Blasendurchmesser, die Blasengrößenverteilung, die Flüssigkeitsgeschwindigkeit und die Gasgeschwindigkeit für zwei verschiedene Turbulenzbedingungen validiert. Anschließend wurden Experimente in einem 3 x 3-Stabbündel mit einem Abstandshalter und Fahnen unter Verwendung der Röntgen-Computertomographie (CT) durchgeführt, die qualitativ hochwertige Gasgehaltdaten liefert, ohne die Strömung zu stören. Die Auswirkungen unterschiedlicher Massen- und Wärmeströme auf den Gasgehalt und seine Verteilung stromabwärts des Abstandshalters wurden analysiert. Außerdem wurden die Auswirkungen verschiedener Fahnenwinkel auf die Verteilung des Gasgehaltes diskutiert. Darüber hinaus wurde eine experimentelle Datenbasis in einem Stabbündel mit einem Abstandshalter unter verschiedenen Strömungsbedingungen gewonnen, um die numerische Modellierung zu validieren. Schließlich wurde das verbesserte CFD-Modell, das aus der halbgeschlossenen Rohrgeometrie gewonnen wurde, auf die 3 x 3 Stabbündelgeometrie bei zwei verschiedenen Turbulenzbedingungen angewendet. Die numerischen Ergebnisse wurden mit den Röntgen-CT-Daten für den Gasgehalt verglichen.

Die vorliegende Arbeit befasst sich mit der Stabilitätsuntersuchung von offenen Naturumlaufsystemen als Grundlage zur Verwendung als passives Wärmeabfuhrsystem im Sicherheitsbehälter eines Siedewasserreaktors. Der Betrieb eines solchen Systems im Naturumlauf basiert einzig auf der Ausbildung eines Dichtegradienten, der infolge einer freien Konvektion zu einer Strömung innerhalb dieses Systems führt. Dieser Dichtegradient im Arbeitsfluid wird durch die Wärmezu- und -abfuhr hervorgerufen. Der sich ausbildende und kontinuierlich steigende Massenstrom geht bei Erreichen der Sättigungstemperatur in Massenstromoszillationen, den sogenannten Zweiphasenströmungsinstabilitäten, über. Mit steigender Temperatur des Arbeitsfluides kehrt der Massenstrom zu einer stabilen Strömung und kontinuierlichen Wärmeabfuhr zurück, jedoch als Zweiphasenströmung. Es wurde an der Technischen Universität Dresden eine Versuchsanlage errichtet, die den Gebäudekondensator des KERENA\textsuperscript{TM}(ehemals SWR1000)-Reaktorkonzepts nachstellt, um die Anlagen- und die Betriebscharakteristik hinsichtlich geometrischer Einflüsse zu bewerten. Mit Hilfe hochauflösender Temperatur und Volumendampfgehaltsmessung wurde festgestellt, dass bei parallel angeordneten Steigrohren die durch Kondensationsschläge hervorgerufenen Druckschläge stark reduziert oder sogar unterbunden werden konnten. So fungiert eines der Steigrohre als Puffer für rückströmendes unterkühltes Fluid aus der Wärmesenke in dem anderen. Zusammengefasst wurde zudem die Betriebscharakteristik in Stabilitätskarten, die die stabile Einphasenströmung, die instabile Zweiphasenströmung und die stabile Zweiphasenströmung eindeutig voneinander abgrenzt. Die Vorhersage der Stabilitätsgrenze zwischen instabiler und stabiler Zweiphasenströmung durch einen analytischen Ansatz ist gelungen. Das zugrundeliegende Modell für ein solches offenes Naturumlaufsystem wurde durch die Methode der gewichteten Residuen und die Finite-Volumen-Methode gelöst und mit Hilfe der Proper Orthogonal Decomposition auf ein Modell niedriger Ordnung reduziert (ROM). Vergleichsrechnungen mit einem entwickelten Abbild des Versuchsstandes GENEVA unter Verwendung des bereits validierten Systemcodes ATHLET der \textit{Gesellschaft für Anlagen- und Reaktorsicherheit (GRS) gGmbH} bestätigten die berechneten Betriebszustände und letztendlich die durch die lineare Stabilitätsuntersuchung ermittelte Stabilitätsgrenze. Eben dieses ROM bildet die Zweiphasenströmung mittels des \textit{Drift-flux mixture} Modells ab, welches die relativen Geschwindigkeiten jeder Phase berücksichtigt. Die nichtlineare Stabilitätsuntersuchung dieses ROMs ergab an ausgewählten Referenzbetriebspunkten superkritische Hopfbifurkationen, die nur durch die Detektion aufkommender stabiler Grenzzyklen während der numerischen Integration nachgewiesen werden konnten. Parameterstudien zur Stabilitätsanalyse können durch dieses ROMs unter erheblicher Reduktion von Rechenaufwand durchgeführt werden.

The majority of technical separation processes for fluid mixtures utilize the principle of rectification. If a two-phase mixture is fed into the column, possibly undesirable flow morphologies or severe droplet carry-over may occur, which detrimentally affect separation efficiency and equipment integrity. Currently, the two-phase flow behavior in feed pipes is hardly predicable and mostly based on empirical or heuristic methods, which do not properly account for a broad range of possible fluid properties and plant dimensions. As a consequence, costly safety margins are applied. Feed pipes to separation columns are often characterized by horizontal inlet nozzles, small length-to-diameter ratios and complex routing, involving elbows or bends. The pipe lengths are too short to enable the two-phase flow to fully develop, which thus, enters the column with unknown flow morphology. Since developing flows have rarely been studied, today’s engineering practice relies on existing predictive methods for fully developed two-phase flows. Graphical methods can hardly represent gradual transitions between flow regimes. Analytical models provide only simplified flow representations of the two-phase flow that have not yet been qualified for developing pipe flow. In this work, a comprehensive experimental database of horizontal water-air flows in two test sections with nominal pipe diameters of D = 50 mm and D = 200 mm and feed pipe lengths in the range 10 < L/D < 75 was established. This way, the data cover developing pipe flows with entrance lengths typical for two-phase feeds of separation columns and more developed flows that are comparable with the extensively studied reference system water-air. A particular focus was put on the effect of pipe bends on the flow morphology up- and downstream. The flow morphology was captured using imaging wire-mesh sensors. A 4D fuzzy algorithm was applied to objectively identify the flow two-phase morphologies. Based on their fuzzy representation, the flow morphologies were classified and a novel 2D visualization technique is proposed to discuss the flow development along the feed pipes. Undesired flow morphologies (intermittent flow and entrainment) during the operation of two-phase feeds are hardly predictable by conventional design tools. The inception of intermittent flows was analyzed using the experimental data. Consequently, the inception criteria based on the required liquid levels for fully developed intermittent flows were adapted for short entrance lengths. The characteristic dynamics of flow morphologies that are known to cause the onset of entrainment were analyzed. Based on wave frequencies, a predictive criterion for the susceptibility of wavy flows for the onset of entrainment is introduced and applied to straight feed pipes and horizontal 90° bends. Among the dozens available, 66 reduced-order models for the prediction of the void fraction were tested for straight feed pipes and horizontal 90° pipe bends. Thereof, the ones most suitable for variable operating conditions and pipe geometries were identified and adapted. Complementary 3D simulations were performed to verify the applicability of numerical codes (VoF, AIAD) for flows with free interfaces. The flow morphologies were successfully reproduced at macroscopic scale, however, the simulation results rank behind reduced-order models considering their quantitative predicting capabilities.:Abstract II Kurzfassung IV Acknowledgement VI Nomenclature VIII Table of Contents XIII 1 Introduction 1 1.1 Thermal separation in view of the 21st century 1 1.2 Engineering and design of rectification plants 2 1.3 Outline of the thesis 4 2 State of the art 5 2.1 Two-phase feeds in thermal separation 5 2.1.1 Feed condition as adjustable parameter 5 2.1.2 Thermohydraulic optimization 8 2.1.3 Hydrodynamic conditioning 9 2.2 Hydrodynamics of two-phase feeds 11 2.2.1 Flow morphologies in feed pipes 11 2.2.2 Droplet entrainment 14 2.2.3 Flow regime maps 17 2.2.4 Consequences for two-phase feeds 19 2.3 Modelling of two-phase feeds 23 2.3.1 Basic definitions 23 2.3.2 Fundamentals of the two-fluid model 25 2.3.3 The interfacial level gradient 29 2.3.4 Analytical models 32 2.3.5 CFD simulations for commercial feed pipes 34 2.4 Objectives of this thesis 36 3 Experimental method and algorithms for flow characterization 37 3.1 Experimental setups 37 3.2 Wire-mesh sensors 40 3.3 Experimental procedure 42 3.4 Data processing 44 3.4.1 Fuzzy flow morphology classification 45 3.4.2 Power spectral density 48 3.5 Measurement uncertainty 49 4 Flow morphologies in different feed pipe geometries 53 4.1 Developing two-phase flow in straight pipes 53 4.2 Effect of pipe curvatures on the flow morphology 55 4.3 Morphology recovery 57 4.4 Conclusions 60 5 Prediction of undesirable flow morphologies in feed pipes 61 5.1 Initiation of intermittent flows 61 5.2 Onset of droplet entrainment 62 5.2.1 Vulnerable flow morphologies 62 5.2.2 Derivation of a criterion for onset of entrainment 64 5.2.3 Adjustment of the criterion for the investigated pipe geometries 67 5.3 Conclusions 70 6 Reduced-order modelling of two-phase feeds 71 6.1 Prediction of void fraction 71 6.2 Liquid levels 75 6.3 Conclusions 78 7 CFD modelling of two-phase feeds 79 7.1 Simulation setup 79 7.2 Multiphase models 82 7.3 Comparison with experimental data 83 7.3.1 Straight pipes 83 7.3.2 Horizontal 90° bends 85 7.4 Conclusions 88 8 Summary and recommendations for future work 89 8.1 Summary 89 8.2 Recommendations for future work 91 References 94 List of figures 113 List of tables 118 Appendix i Scientific publications and contributions xxxiii Eidesstattliche Erklärung xxxvii / Die meisten technischen Verfahren zur Trennung von Flüssigkeitsgemischen beruhen auf dem Prinzip der Rektifikation. Wird ein Zweiphasengemisch in die Trennkolonne eingespeist, können unerwünschte Strömungsmorphologien oder ausgeprägte Tröpfchenverschleppung auftreten, welche sich nachteilig auf die Trennleistung und die Integrität einzelner Anlagenkomponenten auswirken. Derzeit lässt sich das Verhalten solcher Zweiphasenströmungen in Einspeiseleitungen kaum vorhersagen und basiert meist auf empirischen oder heuristischen Methoden, die ein breites Spektrum möglicher Stoffeigenschaften und Anlagendimensionen nicht angemessen berücksichtigen. Infolgedessen müssen kostspielige Sicherheitszuschläge angewendet werden. Einspeiseleitungen von Trennkolonnen sind häufig durch horizontale Eintrittsstutzen, ein geringes Länge-zu-Durchmesser-Verhältnis und eine komplexe Leitungsführung mit Bögen und anderen Normteilen gekennzeichnet. Typische Rohrlängen sind zu kurz, um eine vollständig entwickelte Zweiphasenströmung auszubilden, welche daher mit unbekannter Strömungs-morphologie in die Trennkolonne eintritt. Da derartige Strömungen jedoch bisher nur selten untersucht wurden, verlässt man sich gegenwärtig in der technischen Praxis auf bestehende Vorhersagemethoden für voll entwickelte Zweiphasenströmungen. Grafische Methoden können jedoch die allmählichen Übergänge zwischen Strömungsformen kaum darstellen. Analytische Modelle liefern nur vereinfachte Näherungswerte der Zweiphasenströmung, die noch nicht für sich entwickelnde Rohrströmung qualifiziert wird. In dieser Arbeit wurde eine umfangreiche experimentelle Datenbasis horizontaler Wasser-Luft-Strömungen in zwei Versuchsstrecken mit Rohrinnendurchmessern von D = 50 mm und D = 200 mm und Einlauflängen im Bereich 10 < L/D < 75 erstellt. Auf diese Weise decken die Daten sowohl sich entwickelnde Rohrströmungen mit typischen Einlauflängen für Einspeiseleitungen ab, als auch weiter (in axialer Richtung) entwickelte Strömungen, die mit dem umfangreich untersuchten Referenzsystem Wasser-Luft vergleichbar sind. Die Auswirkung von Rohrbögen auf die Strömungsmorphologie stromauf- und stromabwärts wurde gezielt untersucht. Die Strömungsmorphologie wurde mit bildgebenden Gittersensoren erfasst. Ein 4D-Fuzzy-Algorithmus wurde zur objektiven Identifizierung der Strömungsmorphologien eingesetzt. Auf Grundlage dieser Fuzzy-Darstellung der Strömung wurden die Strömungsmorphologien klassifiziert, und es wurde eine neuartige 2D-Visualisierungstechnik entworfen, mit der die Strömungsentwicklung entlang der Einspeiseleitungen diskutiert wurde. Unerwünschte Strömungsmorphologien (intermittierende Strömung und Tropfenmitriss) während des Betriebs zweiphasiger Einspeisungen sind mit herkömmlichen Auslegungswerkzeugen kaum vorherzusagen. Das Einsetzen intermittierender Strömungen wurde auf Grundlage der experimentellen Daten analysiert. Daraufhin wurden existierende Kriterien, basierend auf den notwendigen Mindestfüllständen, für das Einsetzen intermittierender Strömungen in Abhängigkeit von den untersuchten Einlauflängen angepasst. Die charakteristische Dynamik von Strömungsmorphologien, die Tropfenmittriss hervorrufen, wurde analysiert. Voraussagemethoden zur Vorhersage der Anfälligkeit welliger Strömungen für das Auftreten von Tropfenmitriss wurden auf der Grundlage von Wellenfrequenzen entwickelt und für gerade Einspeiserohre und horizontale 90°-Bögen angewandt. Von den zahlreichen verfügbaren Modellen zur Vorhersage des Gasanteils wurden 66 Modelle reduzierter Ordnung für gerade Einspeiseleitungen und horizontale 90°-Rohrbögen getestet. Davon wurden die für variable Betriebsbedingungen und Rohrgeometrien am besten geeigneten Modelle ermittelt und angepasst. Komplementäre 3D-Simulationen wurden durchgeführt, um die Anwendbarkeit numerischer Codes (VoF, AIAD) für Strömungen mit freien Grenzflächen zu bestätigen. Die Strömungsmorphologien wurden im makroskopischen Maßstab erfolgreich reproduziert, die Simulationsergebnisse bleiben jedoch hinsichtlich ihrer quantitativen Vorhersagekraft hinter den Modellen reduzierter Ordnung zurück.:Abstract II Kurzfassung IV Acknowledgement VI Nomenclature VIII Table of Contents XIII 1 Introduction 1 1.1 Thermal separation in view of the 21st century 1 1.2 Engineering and design of rectification plants 2 1.3 Outline of the thesis 4 2 State of the art 5 2.1 Two-phase feeds in thermal separation 5 2.1.1 Feed condition as adjustable parameter 5 2.1.2 Thermohydraulic optimization 8 2.1.3 Hydrodynamic conditioning 9 2.2 Hydrodynamics of two-phase feeds 11 2.2.1 Flow morphologies in feed pipes 11 2.2.2 Droplet entrainment 14 2.2.3 Flow regime maps 17 2.2.4 Consequences for two-phase feeds 19 2.3 Modelling of two-phase feeds 23 2.3.1 Basic definitions 23 2.3.2 Fundamentals of the two-fluid model 25 2.3.3 The interfacial level gradient 29 2.3.4 Analytical models 32 2.3.5 CFD simulations for commercial feed pipes 34 2.4 Objectives of this thesis 36 3 Experimental method and algorithms for flow characterization 37 3.1 Experimental setups 37 3.2 Wire-mesh sensors 40 3.3 Experimental procedure 42 3.4 Data processing 44 3.4.1 Fuzzy flow morphology classification 45 3.4.2 Power spectral density 48 3.5 Measurement uncertainty 49 4 Flow morphologies in different feed pipe geometries 53 4.1 Developing two-phase flow in straight pipes 53 4.2 Effect of pipe curvatures on the flow morphology 55 4.3 Morphology recovery 57 4.4 Conclusions 60 5 Prediction of undesirable flow morphologies in feed pipes 61 5.1 Initiation of intermittent flows 61 5.2 Onset of droplet entrainment 62 5.2.1 Vulnerable flow morphologies 62 5.2.2 Derivation of a criterion for onset of entrainment 64 5.2.3 Adjustment of the criterion for the investigated pipe geometries 67 5.3 Conclusions 70 6 Reduced-order modelling of two-phase feeds 71 6.1 Prediction of void fraction 71 6.2 Liquid levels 75 6.3 Conclusions 78 7 CFD modelling of two-phase feeds 79 7.1 Simulation setup 79 7.2 Multiphase models 82 7.3 Comparison with experimental data 83 7.3.1 Straight pipes 83 7.3.2 Horizontal 90° bends 85 7.4 Conclusions 88 8 Summary and recommendations for future work 89 8.1 Summary 89 8.2 Recommendations for future work 91 References 94 List of figures 113 List of tables 118 Appendix i Scientific publications and contributions xxxiii Eidesstattliche Erklärung xxxvii

[pt] A ocorrência do escoamento slug em tubulações horizontais é de especial interesse para a indústria de petróleo devido aos riscos operacionais indesejados associados a esse padrão de escoamento. Portanto, nas últimas décadas um intenso esforço foi dedicado ao estudo e modelagem do escoamento slug. Ferramentas preditivas baseadas na estabilidade linear de Kelvin-Helmhotz foram amplamente desenvolvidas na literatura para prever a transição para esse regime de escoamento. Esses modelos são derivados da análise de estabilidade modal de perturbações bem definidas. No entanto, para escoamento em tubulação, um número bastante limitado de estudos experimentais dedicados para investigação da evolução de perturbações que originem o regime slug está disponível. Além disso, estudos a partir da introdução de perturbações bem definidas, que podem fornecer informações precisas para validação de modelos e simulações numéricas, foram encontrados. O presente trabalho abordou o problema da transição para o regime slug a partir da caracterização da evolução de ondas interfaciais. Essas perturbações controladas foram excitadas com um modo de geração na interface do escoamento estratificado utilizando uma placa oscilatória. O trabalho se concentra na caracterização de ondas interfaciais no regime linear, que corresponde ao regime de estudo da maioria dos modelos disponíveis na literatura. Portanto, um limiar de amplitude para ondas lineares foi estimado experimentalmente. O acionamento da placa oscilatória foi sincronizado com as aquisições de imagens, permitindo medições sincronizadas em fase. As medições do campo de velocidade foram realizadas usando a técnica de Velocimetria de Imagem de Partículas (PIV) e Iluminação de Fundo (Shadowgraphy). O perfil de velocidade e turbulência do escoamento foram medidos simultaneamente nas fases do liquido e do gás. A sincronização em fase permitiu a extração do perfil de flutuação de velocidade coerentes as ondas interfaciais. Os resultados obtidos são originais e mostraram, pela primeira vez na literatura, que os modos interfaciais em ambas as fases são quase independentes dos modos cisalhantes, dentro da faixa de parâmetros abordados neste trabalho. A caracterização de ondas não lineares foi brevemente investigada, indicando mudanças no perfil do escoamento médio. Além disso, foi obtida uma correlação para o fator de atrito das ondas interfaciais, levando a uma melhoria na estimativa da altura do líquido e da perda de carga do tubo quando combinadas nas relações de fechamento dos modelos 1-D. A metodologia experimental proposta neste trabalho é uma ferramenta valiosa para produzir informações precisas que podem ser usadas para validar e aprimorar modelos teóricos e simulações numéricas. O estudo pode contribuir para a compreensão dos mecanismos físicos envolvidos na transição do escoamento estratificado para slug. / [en] The occurrence of slug regime in horizontal pipelines is of special interest for the oil and gas industry due to the unwanted operational risks associated with this flow. Hence, an intense effort has been devoted to the study and to model this flow regime. Predictive tools based on linear Kelvin-Helmhotz stability have been widely applied in the literature for prediction of slug onset. These models are derived from stability analysis of well-defined disturbances. However, for pipe flows, a limited number of experimental studies devoted to investigate the evolution of disturbances that lead to the initiation of slugs is available. In addition, no studies are found using of well define disturbances, which could provide accurate information for validation of models and numerical simulations. The present work addresses the problem by the studying of the evolution of controlled waves excited at the liquid interface. To this end, an oscillating paddle was employed. The work focuses the characterization of interfacial waves within the linear regime, which correspond to the regime of most models available in the literature. The amplitude threshold for linear waves was experimentally estimated. The driving signal of the oscillating paddle was synchronized with image acquisitions, enabling phase locked measurements of the waves and hence the use of ensemble averaging techniques. Phase-locked measurements of the velocity field in the liquid and gas layers were performed using off-axis Particle Image Velocimetry (PIV) technique and Shadowgraph. Mean flow, streamwise and wall normal fluctuations were measured simultaneously in the liquid and gas phases. For a range of flow rates and exciting wave frequencies the combined techniques employed allowed the extraction from the measured velocity fields, the coherent part of flow fluctuations related with the exciting waves. The results obtained have shown, seemingly, for the first time, that interfacial modes in both phases are nearly independent of near wall disturbances within the range of parameters covered in this work. Characterization of nonlinear waves was briefly investigated indicating changes in the mean velocity. Moreover, a correlation for wave friction factor based on wave and flow parameters was obtained, leading to an improvement on the liquid heightand pipe head loss estimation when are combined into the closure relations used for the 1-D models. The experimental methodology proposed in this work is a valuable tool to produce accurate information that can be used to validate and improve theoretical models and numerical simulations. It can contribute to the understanding of the physical mechanisms involved in the transition from stratified to slug flows.

<p>A short, 30cm, test section was used to study the effect of electrohydrodynamic (EHD) forces on flow redistribution in a horizontal, shell and tube heat exchanger subject to both boiling and condensation. The use of a short test section allows for a consistent flow pattern across the test section length which provides further insight into the true effect of EHD.</p> <p>It was found that the voltage polarity of the applied voltages influences the flow distribution. For the current geometry studied, it was found that positive polarity voltages tend to pull liquid away from heat transfer surface and that negative voltages tended to repel more liquid toward the heat transfer surface. Using this knowledge we were able to show that positive voltages were more effective for convective condensation heat transfer enhancement, whereas negative voltages were more effective for convective boiling heat transfer enhancement. A twofold enhancement of convective boiling heat transfer was achieved for positive voltages and a 4fold enhancement was achieved for negative voltages. Similar pressure drop penalties were seen for both cases, approximately twice that of the no EHD case.</p> <p>Furthermore, the effect of DC level, peak to peak voltage, frequency and duty cycle waveform parameters on convective boiling enhancement were studied to explore the range of controllability for the current set of flow parameters. It was found that these various waveform parameters can induce different flow patterns and consequently different heat transfer and pressure drop configurations. In general the heat transfer is enhanced by EHD, but different pressure drop penalties can be achieved for a given enhancement ratio using different waveforms. High heat transfer for relatively low pressure drop was achieved using either negative DC signals or 50%duty cycle pulse waveforms. In some cases the enhancement is quite little compared to the pressure drop, for example the zero DC level, varying peak to peak voltage data. It is suggested that in a system where the heat exchanger pressure drop due to EHD is more dominant than the system pressure drop, it may be possible to use EHD as a method of retarding the system rather than enhancing it thereby broadening the scope of controllability.</p> <p>Finally we showed the proof of concept of using DC EHD as a rapid control mechanism for the load conditions. Using -8kVDC the water side heat flux could be varied by approximately ±3.2 kW/m² within 5 seconds. As a comparison, the same experiment was repeated using the refrigerant flow rate to control the load. Response times were similar for both experiments and although the power required for the flow rate control was less, the minimal variability in flow parameters for the EHD control make it a more attractive method of load control.</p> / Master of Applied Science (MASc)

<p>Flow-Accelerated Corrosion (FAC) is a pipe wall thinning mechanism affecting carbon steel piping systems in power generation plants. Mass transfer is the rate limiting factor, even though chemistry and materials determine the overall potential for FAC. Different localized thinning rates in back to back elbow configurations between the first and second elbow have been noted at nuclear power plants, and this difference depends on the length of pipe between the elbows, flow conditions, and the configuration of the back to back elbows (e.g. S, C, or out of plane). In this thesis, mass transfer measurements in back to back elbows arranged in an out of plane configuration under single and annular two-phase flow conditions are presented.</p> <p>The mass transfer measurements were performed using a wall dissolving technique with bend sections cast from gypsum. The diffusivity of gypsum in water is similar to the diffusivity of iron from the magnetite layer of carbon steel pipe in water, thus providing analogous mass transfer conditions to FAC in power generation plants. The wall dissolution of gypsum allows the surface roughness to develop due to the flow. The mass transfer is determined by passing water through the gypsum test sections in a flow loop system. The test sections are then sectioned into two halves to expose the worn surface. The surface topology is measured using a three dimensional laser scanner. The wear progression of the surface with time provides local mass transfer rates, locations of high mass transfer and local surface roughness.</p> <p>The single-phase flow experiments were performed at a Reynolds number of 70,000 for different lengths of pipe (0, 1, 2 and 5 pipe diameters) between the elbows. The mass transfer results show regions of higher mass transfer in the second elbow in comparison to the first el­­bow. The maximum mass transfer rate in the second elbow decreases when the length of the pipe between the elbows was increased from 0 to 5 pipe diameters. Surface features corresponding to flow streaks on the second elbow surface indicated swirling flow, and its strength decreases with increasing separation distance between the elbows. The surface roughness was found to be higher in the regions of high mass transfer and decreases with increasing elbow separation distance.</p> <p>The effect of air and water superficial velocities on the mass transfer for the bends with a separation distance of 0 pipe diameters was measured under two-phase air-water annular flow. In addition, the effect of separation distance of 0, 1 and 5 pipe diameters in length between the elbows was studied for one annular flow condition. The highest mass transfer was found on the outer wall of the first elbow for all cases. The maximum mass transfer in the second elbow was found to be approximately 60 percent of the maximum value in the first elbow, and was not affected significantly when the elbow separation distance was increased from 0 to 1 and 5 pipe diameters. The separation distance between the elbows did not affect the maximum mass transfer on the outer wall of the first elbow. The mass transfer increased with an increase in either the water or air superficial velocity, with the air velocity having a greater effect. The mass transfer enhancement factor relative to that in a straight pipe only increases significantly with increasing air superficial velocity. The roughness development in the pipe was modest, but increases significantly in the high mass transfer region of the first and second elbow.</p> / Master of Applied Science (MASc)

En este estudio se han construido dos diferentes instalaciones para investigar primero los chorros de gas horizontales y en segundo lugar los chorros verticales de agua que impactan sobre superficies libres de fluido, también se ha desarrollado un modelo numérico integral para predecir las trayectorias de estos jets y sus parámetros más importantes, validándose con los resultados experimentales obtenidos. En la primera parte de este trabajo, se han realizado experimentos para investigar el comportamiento de chorros de gas horizontales penetrando en agua. Los resultados experimentales indicaron que la longitud de penetración de los chorros de gas está fuertemente influenciada por el diámetro de la boquilla y el número de Froude, así como con el flujo de masa de de entrada y su momento. Aumentar el número de Froude y el diámetro del inyector lleva a aumentar la inestabilidad de jet. Además, la máxima ubicación antes de jet pinch-off se muestra que mantiene una relación logarítmica con el número de Froude para todos los diámetros de jet. Se han desarrollado correlaciones empíricas para predecir estos parámetros. Se ha desarrollado un modelo basado en la integración de las ecuaciones de conservación para que resulte útil en el diseño de aplicaciones en las que participen chorros horizontales así como para asistir a la investigación experimental. Las predicciones del modelo integral se comparan con los datos de los datos experimentales obtenidos con muy buenos resultados. En la segunda parte de este trabajo, se realizaron una serie de experimentos con de chorros de agua, inyectados verticalmente hacia abajo, a través de toberas circulares que impactan sobre una superficie de agua. Los resultados obtenidos mostraron que la profundidad de penetración de la burbuja disminuye con la longitud del chorro, pero que después de ciertas condiciones se mantiene casi constante. Además ésta aumenta con los diámetros de la boquilla y la velocidad del chorro. La velocidad de arrastre / Harby Mohamed Abd Alaal, K. (2012). Experimental and Theoretical Study of the Characteristics of Submerged Horizontal Gas Jets and Vertical Plunging Water Jets in Water Ambient [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/18065