Cette thèse est motivée par le besoin de compléter les connaissances actuelles des phénomènes ayant lieu lors d’un renvoi d’eau dans un lit de débris radioactifs, opération appelée « renoyage » et qui intervient dans une séquence d’accident grave où un cœur de réacteur nucléaire est dégradé suite à une perte prolongée de refroidissement primaire. Notre étude, de nature expérimentale, vise à mieux caractériser la crise d’ébullition en convection forcée, dans un milieu poreux chauffant. Le cœur du dispositif expérimental est un milieu poreux modèle quasibidimensionnel, composé de 276 cylindres disposés entre deux plaques de céramique distantes de 3 mm, dont l’une, transparente, permet de visualiser les écoulements. Les cylindres, de 2 mm de diamètre, sont des sondes thermo-résistives qui ont une double fonction : elles sont utilisées comme éléments chauffants et comme capteurs de température. Une boucle fluide permet de contrôler le débit d’injection de liquide dans la section test, la température d’injection ainsi que la pression. La section test est placée verticalement, le liquide est injecté par le bas à une température proche de la saturation. Dans une première série d’expériences, la puissance thermique dissipée globalement par un ensemble de cylindres chauffants est augmentée de façon progressive jusqu’à atteindre l’assèchement d’une zone du milieu poreux. Les résultats montrent deux types de phénoménologies dans le déclenchement de la crise d’ébullition. Pour des débits d’injection faibles (densités de flux massique de l’ordre de 4 kg.m^-2.s^-1 maximum), l’atteinte de la puissance d’assèchement se traduit par un lent recul du front diphasique jusqu’à sa stabilisation en haut de la zone chauffée ; en aval de la zone chauffée, l’écoulement est monophasique vapeur. Pour des débits d’injection plus élevés, la crise d’ébullition apparaît autour d’un des éléments chauffants, conduisant à une ébullition en film localisée, tandis qu’un écoulement diphasique liquide-vapeur continue de parcourir l’aval de la section test. Les visualisations de ces expériences permettent d’identifier qualitativement la structure des écoulements. D’autres expériences consistent à mesurer le flux critique local autour d’un cylindre choisi, pour différentes configurations d’écoulements. Le débit d’injection est fixé. Une puissance de chauffe est imposée à une ligne horizontale de cylindres en amont du cylindre choisi. Les résultats montrent que le flux critique sur ce cylindre diminue en fonction de la puissance délivrée à la ligne chauffée. La distance du cylindre étudié à la ligne chauffée semble avoir peu d’influence sur le flux critique. Des visualisations expérimentales sont utilisées pour caractériser l’écoulement diphasique en aval de la ligne chauffée, dans le but de mettre en relation le flux critique local avec des paramètres hydrodynamiques (saturations, vitesses des phases). Les images obtenues sont difficiles à exploiter. Afin de calibrer les paramètres des algorithmes de traitement d’images, nous avons reproduit une cellule d’essai de géométrie identique à l’originale, mais où l’on injecte du gaz par une ligne de cylindres en amont de la section test dans une configuration d’écoulement diphasique isotherme. Dans ce dispositif, le débit d’injection de gaz est contrôlé et mesuré. Les visualisations obtenues servent alors de références auxquelles sont comparées les visualisations d’ébullition convective. / This work is motivated by the need to better understand the phenomena occurring while some water is injected into a heated porous debris bed. This reflooding operation is a part of the planned mitigation procedure used during a Loss Of Coolant Accident (LOCA) that may occur into a nuclear power plant and results into a severe core damage. Our experimental study aims to characterize the boiling crisis that can happen in a boiling flow taking place within a heatgenerating model porous medium. The test section is a two-dimensional model porous medium, composed of an array of 276 cylinders placed between two ceramic plates spaced from one another by 3 mm, one of which is transparent and allows visualizations of the flow. The 2 mm diameter cylinders are Pt100 resistance temperature detectors that perform a dual function: they act as heating elements (heated by Joule effect) and are also used as temperature probes. A fluid loop allows controlling the liquid injection flow rate, its inlet temperature as well as its pressure. The test section is held vertically, the liquid injected from bottom at a temperature close to the saturation temperature. In a first series of experiments, the thermal power applied to a bundle of heating cylinders is progressively increased until a dry zone is detected in the porous medium. Two kinds of phenomenology are observed during these “dryout experiments”. First, at low liquid injection rate (4 kg.m^-2.s^-1 maximum mass flux), reaching the dryout power results into a liquid front receding down to the upper limit of the heated zone, while downstream the heated zone, the porous medium is vapour-saturated. Second, at higher flow rate, the boiling crisis happens at the surface of a single heating element, resulting in a local film boiling, whereas a two-phase flow still go through the whole test section. High-speed visualizations allow characterizing the flow regimes. Other experiments focus on determining the local critical heat flux on a given cylinder, for different upstream flow configurations. The inlet liquid flow rate is fixed. A thermal power is uniformly applied to a line of heating cylinders, upstream the cylinder under study. Results show that the local critical heat flux decreases as the power applied to the heated line increases. The distance from the cylinder under study to the heated line seems not to have a significant effect on the critical heat flux. Visualizations are used to characterize the two-phase flow upstream the heated line, aiming at expressing the critical heat flux as a function of the hydrodynamic parameters (saturations, phase velocities). The image analysis is particularly challenging. In order to calibrate the image processing parameters, we use a second model porous medium with the same geometry as the heat generating one, but where an isothermal two-phase flow is obtained by injecting gas into the liquid flow rather than generated by boiling. The gas injection flow rate is controlled and measured. Isothermal two-phase flow visualizations provide a reference case and are compared to flow boiling visualizations.

Made available in DSpace on 2014-10-09T12:42:40Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T13:57:16Z (GMT). No. of bitstreams: 1 05031.pdf: 4962096 bytes, checksum: 39c12c06c0063abb20c1c82005ecef33 (MD5) / Tese (Doutoramento) / IPEN/T / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

A pesquisa realizada tratou do estudo da transferência de calor e do fluxo crítico durante a ebulição convectiva no interior de canais de diâmetro reduzidos a partir de dados levantados em bancadas experimentais construídas para esta finalidade. Extensa pesquisa bibliográfica foi efetuada e os principais métodos disponíveis para previsão de coeficiente de transferência de calor, fluxo crítico e mapas de escoamento foram levantados. Os resultados obtidos foram parametricamente analisados e comparados com os métodos da literatura. Pela primeira vez para microcanais, resultados experimentais foram levantados por um mesmo autor em laboratórios distintos buscando verificar a tendência e comportamentos. Tal comparação tem sua importância destacada em face das elevadas discrepâncias observadas na literatura quando resultados de autores distintos, obtidos em condições similares, são comparados. Os resultados levantados foram utilizados na elaboração de modelos que consideram os padrões de escoamento observados em microcanais. A incorporação dos padrões permitiu o desenvolvimento de modelos mecanísticos para coeficiente de transferência de calor, fluxo crítico e critérios para a caracterização da transição entre macro e microcanais baseados na formação do padrão de escoamento estratificado e na simetria do filme líquido no escoamento anular. / This research comprises an experimental and theoretical study on flow boiling heat transfer and critical heat flux inside small diameter tubes based on data obtained in experimental facilities specially designed for this purpose. A broad literature review was carried out and the main methods to predict the heat transfer coefficient, critical heat flux and flow patterns were pointed out. The experimental results were parametrically analyzed and compared against the predictive methods from literature. For the first time, microchannels experimental results obtained by an unique researcher in distinct laboratories were compared and a reasonable agreement was observed. The importance of such a comparison is high-lighted for flow boiling inside microchannels due to the high discrepancies ob-served when results from independent laboratories obtained under similar experimental conditions are compared. Moreover, the experimental results obtained in the present study were used to develop correlations and models for the heat transfer coefficient and heat flux that takes into account the flow patterns observed in microchannels. The heat transfer coefficient and critical heat flux models were developed based on mechanistic approach. In addition, criteria to characterize macro to microchannel transition were proposed based in the occurrence of the stratified flow pattern and the liquid film symmetry under annular flow conditions.

This thesis deals with the problem of critical heat flux (CHF) on technically smooth and treated surfaces at low pressures. The theoretical part presents the basic concepts of two-phase flow and an analysis of existing work on the influence of the surface on CHF. The main part of the work describes the built experimental apparatus for CHF research at low pressures of 100 -1500 kPa (1-15 bar) with a vertical internally heated annular test section. The internal annuli consists of an outer glass tube with an inner diameter of 14.8 mm and an inner tube made of Inconel ™ 625 / Optimized ZIRLO ™ with an outer diameter of 9.14 mm and a heated length of 380/365 mm. CHF experiments on technically smooth surface were performed at outlet pressures 120 kPa, 200 kPa and 300 kPa, at an inlet temperature of 64, 78 and 91 °C and at mass flux of 400, 500, 600 and 800 kg / m2s. The Inconel tubes were tested in two different surface modifications - abraded and bead blasted. Experiments were performed at mass flows of 400, 500 and 600 kg / m2s. The total number of 122 experimental runs were conducted and the results were compared with other literature experimental data. The maximum increase of CHF on abraded / bead blasted tube was 18.12% / 16.17%. The surface structure was analysed by laser microscopy. The wetting behaviour of the surface structures was measured by the sessile drop method. The elemental analysis of the surface was evaluated using the EDS method.

A unifying model for Critical Heat Flux (CHF) prediction has been elusive for over 60 years. With the release of the data utilized in the making of the 2006 Groeneveld Lookup table (LUT), by far the largest public CHF database available to date, data-driven predictions on a large variable space can be performed. The popularization of machine learning techniques to solve regression problems allows for deeper and more advanced tools when analyzing the data. We compare three different machine learning algorithms to predict the occurrence of CHF in vertical, uniformly heated round tubes. For each selected algorithm (ν-Support vector regression, Gaussian process regression, and Neural network regression), an optimized hyperparameter set is fitted. The best performing algorithm is the Neural network, which achieves a standard deviation of the prediction/measured factor three times lower than the LUT, while the Gaussian process regression and the ν-Support vector regression both lead to two times lower standard deviation. All algorithms significantly outperform the LUT prediction performance. The neural network model and training methodology are designed to prevent overfitting, which is confirmed by data analysis of the predictions. Additionally, a feasibility study of transfer learning and uncertainty quantification is performed, to investigate potential future applications.

Addition of nano particles to cooling fluids has shown marked improvement in the heat transfer capabilities. Nanofluids, liquids that contain dispersed nanoparticles, are an emerging class of fluids that have great potential in many applications. There is a need to understand the fundamental behavior of nano dispersed particles with respect to their agglomeration characteristics and how it relates to the heat transfer capability. Such an understanding is important for the development and commercialization of nanofluids. In this work, the stability of nano particles was studied by measuring the zeta potential of colloidal particles, particle concentration and size. Two different sizes of silica nano particles, 10 nm and 20 nm are used in this investigation at 0.2 vol. % and 0.5 vol. % concentrations. The measurements were made in deionized (DI) water, buffer solutions at various pH, DI water plus HCl acid solution (acidic pH) and DI water plus NaOH solution (basic pH). The stability or instability of silica dispersions in these solutions was related to the zeta potential of colloidal particles and confirmed by particle sizing measurements and independently by TEM observations. Low zeta potentials resulted in agglomeration as expected and the measured particle size was greater. The heat transfer characteristics of stable or unstable silica dispersions using the above solutions were experimentally determined by measuring heat flux as a function of temperature differential between a nichrome wire and the surrounding fluid. These experiments allowed the determination of the critical heat flux (CHF), which was then related to the dispersion characteristics of the nanosilica in various fluids described above. The thickness of the diffuse layer on nano particles was computed and experimentally confirmed in selected conditions for which there was no agglomeration. As the thickness of the diffuse layer decreased due to the increase in salt content or the ionic content, the electrostatic force of repulsion cease to exist and Van der Waal's force of agglomeration prevailed causing the particles to agglomerate affecting the CHF. The 10nm size silica particle dispersions showed better heat transfer characteristics compared to 20nm dispersion. It was also observed that at low zeta potential values, where agglomeration prevailed in the dispersion, the silica nano particles had a tendency to deposit on the nickel chromium wire used in CHF experiments. The thickness of the deposition was measured and the results show that with a very high deposition, CHF is enhanced due to the porosity on the wire. The 10nm size silica particles show higher CHF compared to 20nm silica particles. In addition, for both 10nm and 20nm silica particles, 0.5 vol. % concentration yielded higher heat transfer compared to 0.2 vol. % concentration. It is believed that although CHF is significantly increased with nano silica containing fluids compared to pure fluids, formation of particle clusters in unstable slurries will lead to detrimental long time performance, compared to that with stable silica dispersions.

<p dir="ltr">Results from the Flow Boiling and Condensation Experiment (FBCE), which collected the first flow boiling and condensation data in long-duration, steady microgravity through experiments performed onboard the International Space Station (ISS), are presented. Prior to the ISS experiments, a new correlation for flow boiling critical heat flux (CHF) is developed from data obtained in Earth gravity at different orientations and short durations of microgravity onboard parabolic flight. The new correlation accounts for the influence of gravity in the direction of the flow, impacting vapor removal from the channel, and perpendicular to the heated walls, affecting bubble detachment from the walls, on flow boiling CHF. Novel flow boiling experiments in long-duration microgravity were performed with one or two opposite walls heated using the Flow Boiling Module (FBM), which simultaneously captures heat transfer data and high speed images of flow patterns. The unique microgravity CHF results are presented, and parametric trends are correlated to variations in flow patterns. The results are divided into subcooled and saturated inlet conditions and applicable correlations are assessed. The newly proposed correlation outperforms is the best preforming for the entire database, validating its use in microgravity. Visual observations leading up to CHF justify use of the Interfacial Lift-off model, which predicts CHF with good accuracy for all operating conditions. The data obtained onboard the ISS is consolidated with the prelaunch database to develop highly accurate artificial neural networks (ANNs) for flow boiling heat transfer and CHF in microgravity. The ANNs are developed using a systematic approach that enables the prediction of physical trends. Instabilities observed during subcooled flow boiling are further investigated in dedicated experiments performed at an elevated data capture rate of 30 Hz and extended image capture period up to 28 s. Criteria was proposed to demarcate the stable and unstable operating conditions, and a new correlation to predict the onset of flow instability is proposed. Lastly, microgravity flow condensation heat transfer experiments were conducted onboard the ISS, yielding the first flow condensation data in stable microgravity. Trends in the data are discussed and the two-phase mixture Reynolds number is found to be strongly correlated to local heat transfer coefficient. A separated flow model for annular flow is found to accurately predict trends in average heat transfer coefficient, but underpredicts the microgravity database.</p>

The objective of this research was to investigate the performance of two-phase spray cooling with HFC-134a and HFO-1234yf refrigerants using practical enhanced heat transfer surfaces. Results of the study were expected to provide a quantitative spray cooling performance comparison with working fluids representing the current and next-generation mobile air conditioning refrigerants, and demonstrate the feasibility of this approach as an alternative active cooling technology for the thermal management of high heat flux power electronics (i.e., IGBTs) in electric-drive vehicles. Potential benefits of two-phase spray cooling include achieving more efficient and reliable operation, as well as compact and lightweight system design that would lead to cost reduction. The experimental work involved testing of four different enhanced boiling surfaces in comparison to a plain reference surface, using a commercial pressure-atomizing spray nozzle at a range of liquid flow rates for each refrigerant to determine the spray cooling performance with respect to heat transfer coefficient (HTC) and critical heat flux (CHF). The heater surfaces were prepared using dual-stage electroplating, brush coating, sanding, and particle blasting, all featuring "practical" room temperature processes that do not require specialized equipment. Based on the obtained results, HFC-134a provided a better heat transfer performance through higher HTC and CHF values compared to HFO-1234yf at all tested surfaces and flow rates. While majority of the tested surfaces provided comparable HTC and modestly higher CHF values compared to the reference surface, one of the enhanced surfaces offered significant heat transfer enhancement.

La méthode de refroidissement basée sur le principe thermosiphon présente un grand intérêt en raison de sa simplicité, de sa nature passive et de son coût faible. Elle est adoptée pour le refroidissement à 4,5 K de l’aimant supraconducteur du détecteur de particules CMS auprès du LHC en construction au CERN à Genève. Le travail présenté dans cette thèse étudie expérimentalement les propriétés thermiques et hydrodynamiques d’un écoulement d’He I diphasique en circulation naturelle. Le dispositif expérimental utilisé consiste en une boucle thermosiphon monobranche composée principalement d’un séparateur de phases, d’un tube descendant et d’une section d’essai. Les expériences ont été réalisées en faisant varier plusieurs paramètres tels que le diamètre des sections d’essai (10 mm ou 14 mm) et le flux de chaleur allant jusqu’à l’apparition de la crise d’ébullition. Ces expériences ont permis de déterminer les lois d’évolution des différentes grandeurs caractérisant l’écoulement (le débit massique de circulation, le débit massique vapeur, le titre massique, le coefficient de friction et le coefficient d’échange thermique) en fonction de la densité du flux de chaleur appliquée. Au regard des résultats obtenus, nous discutons la validité des différents modèles classiques existants dans la littérature. Nous montrons que le modèle homogène est le modèle le mieux adapté pour prédire les propriétés hydrodynamiques de ce type d’écoulement dans la gamme de titre massique 0?x?30%. De plus, nous proposons deux modèles pour la prédiction du coefficient de transfert de chaleur diphasique et la densité de flux de chaleur critique. Le premier considère que les effets de la convection forcée et de l’ébullition nucléée agissent simultanément et contribuent au transfert de chaleur. Le deuxième corrèle la densité de flux de chaleur critique mesurée en fonction du rapport altitude sur diamètre / The method of cooling based on the thermosiphon principle is of great interest because of its simplicity, its passivity and its low cost. It is adopted to cool down to 4,5 K the superconducting magnet of the CMS particles detector of the Large Hadron Collider (LHC) experiment under construction at CERN, Geneva. This work studies heat and mass transfer characteristics of two phase He I in a natural circulation loop. The experimental set-up consists of a thermosiphon single branch loop mainly composed of a phase separator, a downward tube, and a test section. The experiments were conducted with varying several parameters such as the diameter of the test section (10 mm or 14 mm) and the applied heat flux up to the appearance of the boiling crisis. These experiments have permitted to determine the laws of evolution of the various parameters characterizing the flow (circulation mass flow rate, vapour mass flow rate, vapour quality, friction coefficient, two phase heat transfer coefficient and the critical heat flux) as a function of the applied heat flux. On the base of the obtained results, we discuss the validity of the various existing models in the literature. We show that the homogeneous model is the best model to predict the hydrodynamical properties of this type of flow in the vapour quality range 0?x?30%. Moreover, we propose two models for the prediction of the two phase heat transfer coefficient and the density of the critical heat flux. The first one considers that the effects of the forced convection and nucleate boiling act simultaneously and contribute to heat transfer. The second one correlates the measured critical heat flux density with the ratio altitude to diameter

The search for Accident tolerant fuels (ATF) which is the first part of this thesis is currently one of the most actual topics in the field of nuclear fuels. These fuels must be first successfully tested in operational and also accident conditions for their possible inclusion in commercial use. Following part of the thesis specifically focuses on the boiling crisis in nuclear reactors which can damage the nuclear fuel cladding. Therefore, it is necessary to know the critical heat flux value and the departure from nuclear boiling ratio. Calculations which determine critical heal flux value are placed in the practical part of the thesis. Calculations are compared with the data obtained during experiments. The ALTHAMC12 and the other correlations which are based on the previous measurements are used for the computational analysis.