Dricksvatten är en essentiell del av ett hållbart samhälle. Därför är det viktigt att säkerställa säkert dricksvatten genom fungerande vattenreningsverk. En viktig förorening att behandla är NOM, som i sig är ofarligt men som kan producera farliga föroreningar. En teknik som används för behandling av NOM är kolfiltrering. Hur ökad flödeshastighet och ökad kontakttid påverkade kolfiltrens effektivitet undersöktes vid dricksvattenreningsverket Norrvatten. De undersökta parametrarna var partiklar, ultraviolett absorbans vid 254 nm, turbiditet, konduktivitet, fluorescent löst organiskt material, totalt organiskt kol, kemisk syreförbrukning, odlingsbara mikroorganismer och lukt. Tre kolfilter studerades vid olika flödeshastigheter; 190, 220, 250 och 280 L/s under ett dygn var. Samtidigt hade två filter ökad kontakttid på 60 och 76 % under sex veckor, medan ett filter fortsatte med den vanliga flödeshastigheten på 190 L/s. Utgående vatten från filtren analyserades för att se om dessa ändringar hade någon effekt på vattenreningen. Denna preliminära studie fann ingen signifikant effekt på kolfiltreringens rening på grund av ökad flödeshastighet eller uppehållstid. Detta kan vara en indikation på att kolfiltreringen kan hantera en framtida flödesökning och därmed vara en väsentlig del av en framtida expansionen av vattenreningsverket. De tecken som visade på att kolfiltren påverkades av ökningen av flödeshastigheter kunde förklaras av fluktuationer i inkommande vatten och skillnader mellan de olika filtren. I framtiden bör effekten av inkommande vatten studeras i detalj. En mer ingående analys av både inkommande och utgående vatten till kolfiltren bör utföras, där provtagning sker oftare för att bättre förstå fluktuationerna i inkommande föroreningskoncentrationer. Dessutom bör testerna upprepas för att se hur reningen skiljer sig från dag till dag. Hur kolfiltren hanterar ökade flödeshastigheter över längre tidsperioder bör också undersökas vidare. / Drinking water is an essential part of a sustainable society. In the future, the demand for drinking water will increase and contaminants in the water sources are also predicted to increase. Therefore, it is essential to ensure safe drinking water through functioning drinking water treatment plants (DWTPs). One important contaminant to treat is natural organic matter (NOM), which is harmless in itself but can produce harmful products. One technique to use for treating NOM is carbon filters (CFs). The effect of increased flow rate and increased empty bed contact time (EBCT) on the CF efficiency was investigated at a DWTP. The investigated parameters were particles, ultraviolet absorbance at 254 nm, turbidity, conductivity, cultivable microorganisms, fluorescent dissolved organic matter, total organic carbon, chemical oxygen demand, and odour. Three CFs were studied at different flow rates; 190, 220, 250, and 280 L/s for 24 hours each. Additionally, two filters had increased EBCT of 60 and 76 %, while one filter continued with the regular flow rate of 190 L/s for six weeks. Outgoing water from the filters was analysed to see if the change had any effect on the DWTP. This preliminary study did not find any significant effect on the CF treatment caused by increased flow rate or EBCT. This could be an indication that the CFs can handle a future increase in flow rate and thus be an essential part of a future expansion of the DWTP. The indications of CFs being affected by the increase in flow rates for some of the parameters could be explained by fluctuations in incoming water or differences between the separate filters. In the future, a more thorough analysis of both incoming and outgoing water to the CFs should be done, where sampling occurs more frequently to better understand the fluctuations in incoming contaminant concentrations. The measurements should also be repeated to see how the treatment differs from day to day. How the CFs handle increased flow rates over longer time periods should also be investigated further.

Heat sinks are used mainly to take away the excessive heat which are produced in a component. This transfer of heat enables a smooth operation of the system with the heat generating component. The efficiency of heat sink is often dependent on the amount of heat it can take away within the smallest duration of time. Several designs and manufacturing techniques have been developed to improve this performance of heat sinks. This project aims at building a test rig which can be used to test the efficiency of heat sinks. The test rig should be designed to be modular, i.e. it should be able to test heat sinks of different sizes and also adhering to the design requirements. The project started with a broad information search on heat sinks and different testing methods. A system architecture was formulated for this test rig as a beginning stage and to find the different architectural components. The main principal components were selected fulfilling the design requirements. A chiller with pump, a flow meter with controller, temperature and pressure sensors and piping’s including a flexible pipe were the main components of the system created. When the specific components were chosen, the design was embodied and the components were arranged in a compact manner as a SolidEdge CAD-model. After several design iterations, a final design was selected. The experimental test rig was then built in the flow lab. The built experimental test setup is able to be adjusted to install heat sinks of different sizes making it modular in design. Future work to improve the performance of the test rig is also suggested. / Värmesänkor, som exempelvis kylflänsar, används främst för att transportera bort värme som orsakas av förluster i en komponent. Denna värmeöverföring möjliggör en smidig drift av systemet där den värmegenererande komponenten ingår. Värmesänkans effektivitet beror på hur mycket värmeeffekt den kan transportera bort. Flera olika konstruktioner och tillverkningstekniker har utvecklats för att förbättra värmesänkors kylprestanda. Målet med detta projekt är att bygga en testrigg som kan användas för att testa kylflänsars effektivitet. Testriggen ska utformas så att den är modulär, dvs. den ska kunna användas för att testa kylflänsar av olika storlekar. Projektet startade med en bred informationssökning om kylflänsar och olika testmetoder. En systemarkitektur skapades som ett början och för att hitta systemets olika principkomponenter. De viktigaste huvudkomponenterna som uppfyllde designkraven valdes sedan. En kylare med pump, en flödesmätare med regulator, temperatur- och tryckgivare och rörledningar, inklusive ett flexibelt rör var de viktigaste komponenterna i det system som skapades. De specifika komponenterna valdes sedan och de representerades och arrangerades på ett kompakt sätt som en system-CAD-modell i SolidEdge. Efter flera iterationer valdes en slutlig konstruktion, och den experimentella testriggen byggdes sedan i flödeslaboratoriet. Den experimentella testriggen kan justeras för att installera kylflänsar av olika storlekar, vilket innebär att den, i viss mån, är modulär. Framtida arbeten för att förbättra testriggens prestanda föreslås också.

This dissertation describes the development of an intelligent non-invasive thermal energy flow rate sensor for laminar and turbulent pipe flows. Energy flow rate is the thermal energy that is carried by a fluid, for example, in a pipe to heat or cool a space in a building. It can be measured by an energy flow rate sensor which consists of a volume flow rate meter and supply and return fluid temperature sensors to bill the users for their energy usage. A non-invasive, low-cost, and easy to install thermal energy flow rate sensor based on thermal interrogation transient heat flux and temperature measurements has been developed to measure fluid velocity and fluid temperature in pipes. This sensor can be used for different pipe diameters, different pipe materials, and different viscous fluids. The transient measurements are made on the outer surface of a pipe by using a heat flux sensor and a thin-film thermocouple which are covered by a thin-film heater. A one-dimensional transient thermal model is applied before and during activation of the external heater along with a parameter estimation code to provide estimates of the fluid heat transfer coefficient and apparent thermal resistance between the thermocouple and the pipe surface. This dissertation contributes to the sensor's development in three ways. First, a new design is developed by using a single layer of Kapton tape with an adhesive (dielectric material) between the thermocouple foils and the pipe wall to isolate the thermocouple electrically from the pipe surface. This new design gives accurate and reliable estimates of the internal mean fluid temperature without environmental interference. Second, this new sensor design is tested for turbulent pipe flows with two different pipe diameters ( = 25.4 mm and = 12.7 mm) and two different viscous fluids (diesel oil and water). Experiments are completed over a large range of fluid velocity from 0.2 m/s to 5.5 m/s and a range of fluid temperature from 20 ℃ to 50 ℃. The estimated parameters, heat transfer coefficient and apparent thermal resistance, are correlated with the fluid velocity and fluid temperature. This sensor gives a good correlation, repeatability, and sensitivity between the estimated parameters and the fluid velocities with an accurate estimation of the fluid temperatures without environmental interference. Third, this sensor is tested for laminar flow in pipes over a range of fluid velocity from 0.049 m/s to 0.45 m/s and a range of fluid temperature from 20 ℃ to 50 ℃. A new empirical correlation between the estimated parameters and the laminar fluid velocity has been developed. The results show that this sensor gives lower sensitivity and accuracy between the estimated parameters and the fluid velocity and fluid temperature for the laminar flow. / Doctor of Philosophy / Heating or cooling is responsible for approximately 50% of the total energy consumption in a building. Budlings' energy consumption can be measured by energy flow rate sensors (measuring both fluid velocity and fluid temperature). Current energy flow rate sensors are invasive (requiring installation inside the system and disturbing the flow) which create unacceptable risks, such as fluid leaks and damage the equipment. Other energy flow rate sensors based on ultrasonic and electromagnetic technologies are non-invasive which can be installed on the outside of the pipe without disturbing the flow, however, they are expensive to buy, difficult to install, and hard to calibrate. Therefore, developing new sensor techniques is necessary, preferably non-invasive, low-cost, and easy to install. In this dissertation, a new non-invasive, low-cost, and easy to install thermal energy flow rate sensor has been designed, developed, and tested. This thermal sensor is based on transient heat flux and temperature measurements which are made on the outside of a copper pipe surface by using a heat flux sensor and a thermocouple. This sensor is used to estimate the energy consumption by measuring a fluid velocity and a fluid temperature in heating and cooling pipe applications for different pipe diameters, different fluids, and different pipe materials. A parameter estimation code is developed to match the analytical and experimental sensor temperature values and to estimate the unknown system parameters. These parameters are correlated with the fluid velocity and fluid temperature. Experiments are completed over a large range of fluid velocity from 0.049 m/s to 5.5 m/s and a range of fluid temperature from 20℃ to 50℃. The encouraging measurement results show that this sensor gives a good correlation, repeatability, accuracy, and sensitivity between the estimated parameters and the fluid velocities with an accurate estimation of the fluid temperatures to allow calculation of the thermal energy consumption.

Current methods for measuring energy flow rate in a pipe use a variety of invasive sensors, including temperature sensors, turbine flow meters, and vortex shedding devices. These systems are costly to buy and install. A new approach that uses non-invasive sensors that are easy to install and less expensive has been developed. A thermal interrogation method using heat flux and temperature measurements is used. A transient thermal model, lumped capacitance method LCM, before and during activation of an external heater provides estimates of the fluid heat transfer coefficient ℎ and fluid temperature. The major components of the system are a thin-foil thermocouple, a heat flux sensor (PHFS), and a heater. To minimize the thermal contact resistance 𝑅" between the thermocouple thickness and the pipe surface, two thermocouples, welded and parallel, were tested together in the same set-up. Values of heat transfer coefficient ℎ, thermal contact resistance 𝑅", time constant 𝜏, and the water temperature °C, were determined by using a parameter estimation code which depends on the minimum root mean square 𝑅𝑀𝑆 error between the analytical and experimental sensor temperature values. The time for processing data to get the parameter estimation values is from three to four minutes. The experiments were done over a range of flow rates (1.5 gallon/minute to 14.5 gallon/minute). A correlation between the heat transfer coefficient ℎ and the flow rate 𝑄 was done for both the parallel and the welded thermocouples. Overall, the parallel thermocouple is better than the welded thermocouple. The parallel thermocouple gives small average thermal contact resistance 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑅"=0.00001 (𝑚2.°C/𝑊), and consistence values of water temperature and heat transfer coefficient ℎ, with good repeatability and sensitivity. Consequently, a non-invasive energy flow rate meter or (BTU) meter can be used to estimate the flow rate and the fluid temperature in real life. / MS / Today, the measuring energy flow rate, measuring flow rate and the fluid temperature, in a pipe is crucial in many engineering fields. In addition, there has been increased use of energy flow rate meters in the renewable energy system and other applications such as solar thermal and geothermal to estimate the useful thermal energy. Some of the commercial energy flow rate meters are using an invasive sensor, has to be inside the pipe, including turbine flow meter and vortex shedding device. These systems are expensive and difficult to install. A new approach that uses non-invasive sensors, attached on the outside of the pipe, that are easy to install and less expensive has been developed by using the heat flux and temperature measurements. A parameter estimation routine was used to analyze the data which depends on the minimum root mean square 𝑅𝑀𝑆 error between the calculated and experimental temperature values. A correlation between the unknown parameter, heat transfer coefficient (ℎ), and the measured flow rate 𝑄 was done to estimate the flow rate. The results show that the new non-invasive system has good repeatability, 15.45%, high sensitivity, and it is easy to install. Consequently, a non-invasive energy flow rate meter or (BTU) meter can be used to estimate the flow rate and the fluid temperature in real life.

This work presents the development of a novel and non-invasive method that measures fluid flow rate and temperature in pipes. While current non-invasive flow meters are able to measure pipe flow rate, they cannot simultaneously measure the internal temperature of the fluid flow, which limits their widespread application. Moreover, devices that are able to determine flow temperature are primarily intrusive and require constant maintenance, which can shut down operation, resulting in downtime and economic loss. Consequently, non-invasive flow rate and temperature measurement systems are becoming increasingly attractive for a variety of operations, including for use in leak detection, energy metering, energy optimization, and oil and gas production, to name a few. In this work, a new solution method and parameter estimation scheme are developed and deployed to non-invasively determine fluid flow rate and temperature in a pipe. This new method is utilized in conjunction with a sensor-based apparatus--"namely, the Combined Heat Flux and Temperature Sensor (CHFT+), which employs simultaneous heat flux and temperature measurements for non-invasive thermal interrogation (NITI). In this work, the CHFT+ sensor embodiment is referred to as the British Thermal Unit (BTU) Meter. The fluid's flow rate and temperature are determined by estimating the fluid's convection heat transfer coefficient and the sensor-pipe thermal contact resistance. The new solution method and parameter estimation scheme were validated using both simulated and experimental data. The experimental data was validated for accuracy using a commercially available FR1118P10 Inline Flowmeter by Sotera Systems (Fort Wayne, IN) and a ThermaGate sensor by ThermaSENSE Corp. (Roanoke, VA). This study's experimental results displayed excellent agreement with values estimated from the aforementioned methods. Once tested in conjunction with the non-invasive BTU Meter, the proposed solution and parameter estimation scheme displayed an excellent level of validity and reliability in the results. Given the proposed BTU Meter's non-invasive design and experimental results, the developed solution and parameter estimation scheme shows promise for use in a variety of different residential, commercial, and industrial applications. / MS / This work documents the development of a novel and non-invasive method that measures fluid flow rate and temperature in pipes. While current non-invasive flow meters are able to measure pipe flow rate, they cannot simultaneously measure the internal temperature of the fluid flow, which limits their widespread application. Moreover, devices that are able to determine flow temperature are primarily intrusive and require constant maintenance, which can shut down operation, resulting in downtime and economic loss. Consequently, non-invasive flow rate and temperature measurement systems are becoming increasingly attractive for a variety of operations, including for use in leak detection, energy metering, energy optimization, and oil and gas production, to name a few. This paper presents a new method that utilizes a non-invasive British Thermal Unit (BTU) Meter based on Combined Heat Flux and Temperature Sensor (CHFT+) technology to determine fluid flow rate and temperature in pipes. The non-invasive BTU Meter uses thermal interrogation to determine different flow parameters, which are used to determine the fluid flow rate and temperature inside a pipe. The method was tested and validated for accuracy and reliability through simulations and experiments. Given the proposed BTU Meter’s noninvasive design and excellent experimental results, the developed novel sensing method shows promise for use in a variety of different residential, commercial, and industrial applications.

This thesis introduces and describes a new experimental setup for examining the effects of pulsating inflow to a ventilated enclosure. The study aimed to test the hypothesis that a pulsating inflow has potential to improve ventilation quality by reducing the stagnation zones through enhanced mixing. The experimental setup, which was a small-scale, two-dimensional (2D), water-filled room model, was successfully designed and manufactured to be able to capture two-dimensional velocity vectors of the entire field using Particle Image Velocimetry (PIV). Using in-house software, it was possible to conclude that for an increase in pulsation frequency or alternatively in the flow rate, the stagnation zones were reduced in size, the distribution of vortices became more homogeneous over the considered domain, and the number of vortices in all scales had increased. Considering the occupied region, the stagnation zones were moved away in a favorable direction from a mixing point of view. In addition, statistical analysis unveiled that in the far-field occupied region of the room model, stronger eddies were developed that we could expect to give rise to improved mixing. As a fundamental experimental study performed in a 2D, small-scale room model with water as operating fluid, we can logically conclude that the positive effect of enhanced mixing through increasing the flow rate could equally be accomplished through applying a pulsating inflow. In addition, this thesis introduces and describes an experimental setup for study of air flow over a wall-mounted radiator in a mockup of a real room, which has been successfully designed and manufactured. In this experimental study, the airflow over an electric radiator without forced convection, a common room-heating technique, was measured and visualized using the 2D PIV technique. Surface blackening due to particle deposition calls for monitoring in detail the local climate over a heating radiator. One mechanism causing particle deposition is turbophoresis, which occurs when the flow is turbulent. Because turbulence plays a role in particle deposition, it is important to identify where the laminar flow over radiator becomes turbulent. The results from several visualization techniques and PIV measurements indicated that for a room with typical radiator heating, the flow over the radiator became agitated after a dimensionless length, 5.0–6.25, based on the radiator thickness. Surface properties are among the influencing factors in particle deposition; therefore, the geometrical properties of different finishing techniques were investigated experimentally using a structured light 3D scanner that revealed differences in roughness among different surface finishing techniques. To investigate the resistance to airflow along the surface and the turbulence generated by the surfaces, we recorded the boundary layer flow over the surfaces in a special flow rig, which revealed that the types of surface finishing methods differed very little in their resistance and therefore their influence on the deposition velocity is probably small. / Det övergripande syftet med den första studien i avhandlingen var att undersöka hypotesen att ett pulserande inflöde till ett ventilerade utrymme har en potential till att förbättra ventilationens kvalitet genom att minska stagnationszoner och därigenom öka omblandningen. För genomförande av studien byggdes en experimentuppställning i form av en tvådimensionell (2D) småskalig modell av ett ventilerat rum. Strömningsmediet i modellen var vatten. Det tvådimensionella hastighetsfältet registrerades över hela modellen med hjälp av Particle Image Velocimetry (PIV). Vid ett stationärt tillflöde bildas ett stagnationsområde i centrum av rumsmodellen. Vid ett pulserade inflöde genererades sekundära virvlar. Med en egen utvecklad programvara var det möjligt att kvantifiera statistiken hos virvlarna. Det pulserade inflödet gjorde att inom området där det vid stationärt tillflöde fanns en stagnationszon ökade antalet virvlar i alla storlekar och fördelningen av virvlar blev mera homogen än tidigare. Detta kan förväntas ge upphov till förbättrad omblandning. Baserat på en grundläggande experimentell studie utförd i en småskalig tvådimensionell rumsmodell med vatten som strömningsmedium kan vi logiskt dra slutsatsen att ett pulserande tilluftsflöde har en potential att förbättra omblandningen.  I en fortsatt studie i avhandlingen visuliserades och mättes hastighetsfältet och därefter beräknades statistiska värden av exempelvis medelhastighet, standardavvikelse och skjuvspänning hos hastighetsfluktuationerna i luftströmmen över en väggmonterad radiator med 2D-PIV-teknik.  Bakgrunden till studien är att en bidragande orsak till partikelavsättning på väggytor är turbofores som uppträder vid en turbulent luftström. Studien genomfördes genom uppbyggnad av en fullskalig rumsmodell. Eftersom turbulens spelar en roll vid partikelavsättning genom turbofores är det viktigt att identifiera var det laminära flödet över radiatorn blir turbulent. Resultaten baserat på visualisering och PIV-mätningar indikerade att, för ett rum med denna typ av radiatoruppvärmning, blev flödet över radiatorn turbulent efter en dimensionslös längd lika med 5,0‒6,25 gånger radiatorns tjocklek. Ytors egenskaper är viktiga vid partikelavsättning. Därför har de geometriska egenskaperna hos några olika metoder för ytbehandling undersökts experimentellt med hjälp av en scanner för strukturerat 3D-ljus. Resultaten visar på skillnader i ytråhet hos de olika ytbehandlingsmetoderna. För att undersöka motståndet mot luftströmning längs ytan och den turbulens som genereras av ytorna registrerade vi gränsskiktsflödet över ytorna i en speciell luftströmningsrigg. Detta påvisade att motståndet hos de olika typerna av ytbehandlingsmetoder skilde sig mycket litet åt och därför är troligt vid deras påverkan på depositionshastigheten mycket liten. / <p>QC 20150525</p>

Nous étudions la faisabilité du procédé de distillation extractive continue pour séparer des mélanges azéotropiques A-B à température de bulle minimale ou maximale, avec un tiers corps E lourd ou léger. Les mélanges ternaires A-B-E appartiennent aux classes 1.0-1-a et 1.0-2 qui se subdivisent chacune en deux souscas selon la position de la courbe d'univolatilité. La colonne de distillation a trois sections, rectification, extractive, épuisement. Nous établissons les équations décrivant les profiles de composition liquide dans chaque section en fonction des paramètres opératoires: pureté et taux de récupération du distillat, taux de reflux ratio R et rapport des débits d'alimentation FE/F dans le cas d'un tiers corps lourd ; pureté et taux de récupération du produit de pied, taux de rebouillage S et rapport des débits d'alimentation FE/F dans le cas d'un tiers corps léger. Avec un tiers corps lourd alimenté comme liquide bouillant au dessus de l'étage d'alimentation du mélange A-B, nous identifions le distillat atteignable et les plages de valeurs faisables des paramètres R et FE/F à partir du critère général de faisabilité énoncé par Rodriguez-Donis et al. (Ind. Eng. Chem. Res, 2009, 48(7), 3544–3559). Pour la classe 1.0-1a, il existe des rapport FE/F et reflux ratio minimum. Le rapport FE/F est plus important pour le procédé continu que pour le procédé discontinu parce que la faisabilité du procédé continu nécessite que les profils d'épuisement et extractifs s'intersectent. Pour la classe 1.0-2, les deux constituants A et B sont des distillats potentiels, l'un sous réserve que le rapport FE/F reste inférieur à une valeur limite maximale. Le procédé continu exhibe également une valeur minimale de FE/F à un taux de reflux ratio donné, contrairement au procédé discontinu. Avec un tiers corps léger alimenté comme vapeur saturante sous l'étage d'alimentation du mélange A-B, nous identifions le produit de pied atteignable et les plages de valeurs faisables des paramètres S et FE/F à partir du critère général de faisabilité énoncé par Rodriguez-Donis et al. (Ind. Eng. Chem. Res, 2012, 51, 4643–4660). Comparé au cas des tiers corps lourds, le produit principal est obtenu en pied. Autrement, les comportements des classes 1.0-1a et 1.0-2 sont analogues entre les tiers corps léger et lourd. Avec un tiers corps léger, le procédé continu ajoute la contrainte que les profils de rectification et extractifs s'intersectent. La contrainte d'intersection des profils d'épuisement et extractif est partagée par les deux modes opératoires continu et discontinu. Ce travail valide la méthodologie proposée pour évaluer la faisabilité du procédé de distillation extractive continue et permet de comparer les tiers entre eux en termes de taux de reflux ratio minimum et de rapport de débit d'alimentation minimal / We study the continuous extractive distillation of minimum and maximum boiling azeotropic mixtures A-B with a heavy or a light entrainer E, intending to assess its feasibility based on thermodynamic insights. The ternary mixtures belong to the common 1.0-1a and 1.0-2 class ternary diagrams, each with two sub-cases depending on the univolatility line location. The column has three sections, rectifying, extractive and stripping. Differential equations are derived for each section composition, depending on operating parameters: distillate product purity and recovery, reflux ratio R and entrainer – feed flow rate ratio FE/F for the heavy case; bottom product purity and recovery, reboil ratio and entrainer – feed flow rate ratio for the light entrainer case. For the case with a heavy entrainer fed as a boiling liquid above the main feed, the feasible product and operating parameters R and FE/F ranges are assessed under infinite reflux ratio conditions by using the general feasibility criterion enounced by Rodriguez-Donis et al. (Ind. Eng. Chem. Res, 2009, 48(7), 3544–3559). For the 1.0-1a class, there exists a minimum entrainer - feed flow rate ratio to recover the product, and also a minimum reflux ratio. The minimum entrainer - feed flow rate ratio is higher for the continuous process than for the batch because of the additional requirement in continuous mode that the stripping profile intersects with the extractive profile. For the 1.0-2 class both A and B can be distillated. For one of them there exists a maximum entrainer - feed flow rate ratio. The continuous process also has a minimum entrainer - feed flow rate ratio limit for a given feasible reflux ratio. For the case with a light entrainer fed as saturated vapor below the main feed, the feasible product and operating parameters S and FE/F ranges are assessed under infinite reflux ratio conditions by using the general feasibility criterion enounced by Rodriguez-Donis et al. (Ind. Eng. Chem. Res, 2012, 51, 4643–4660), Compared to the heavy entrainer case, the main product is removed from the column bottom. Similar results are obtained for the 1.0-1a and 1.0-2 class mixtures whether the entrainer is light or heavy. With a light entrainer, the batch insight about the process feasibility holds for the stripping and extractive sections. Now, an additional constraint in continuous mode comes from the necessary intersection between the rectifying and the extractive sections. This work validates the proposed methodology for assessing the feasibility of continuous extractive distillation processes and enables to compare entrainers in terms of minimum reflux ratio and minimum entrainer feed flow rate ratio

La diminution de la cylindrée ou le downsizing du moteur est potentiellement l'une des stratégies les plus efficaces pour améliorer la consommation de carburant et diminuer les émissions polluantes. Dans le domaine de la suralimentation, la simulation est limitée par les caractéristiques de fonctionnement des turbines fournies par les constructeurs. Une extrapolation précise et fiable des cartographies turbine est donc l’objectif de cette thèse. Une étude expérimentale sur une turbine radiale d’un turbocompresseur est effectuée avec différentes techniques pour mesurer la cartographie turbine la plus large possible. Les mesures sont effectuées sur un banc turbocompresseur classique avec différentes températures d'entrée turbine. Puis une technique de gavage en entrée et en sortie compresseur est testée. Le compresseur est ensuite remplacé par un autre compresseur à roue inversée qui peut aider la turbine à tourner et même l’entrainer. Les débits les plus faibles et même les débits négatifs sont mesurés. Un banc turbine électromécanique a également été développé, mais n’a pas pu donner de résultats satisfaisants à cause de problèmes techniques mais des évolutions à venir restent prometteuses. Les diverses techniques expérimentales testées ont aussi permis de mesurer le rendement isentropique de la turbine et le rendement mécanique du turbocompresseur. Finalement, plusieurs modèles d’extrapolation des courbes caractéristiques turbine ont été testés et confrontés aux résultats expérimentaux. / Engine downsizing is potentially one of the most effective strategies being explored to improve fuel economy and reduce emissions. In the field of turbocharging,simulation is limited by the operating characteristics of turbines supplied by the manufacturers. An accurate and precise extrapolation of the turbine performance maps is the main aim of this study. An experimental study was done on a radial turbine of a turbocharger with different techniques to measure the wider turbine performance map possible. Measurements were done on a classic turbocharger test bench with different turbine inlet temperatures. Then air was blown to the compressor inlet and exit: it is the compressor “gavage”. The compressor is then replaced with another one with are versed rotor: this compressor can help the turbine turn and even drive it itself. The lowest mass flow rates are measured even the negative ones. An electromechanical turbine test bench was developed but did not work correctly because of technical problems but future developments are promising. The various experimental techniques used allowed also the measurement of the turbine isentropic efficiency and the turbocharger mechanical efficiency. Finally, many extrapolation models of the turbine performance maps were tested and compared to the experimental results.

L’étude du transfert de masse turbulent aux interfaces gaz-liquide est d’un grand intérêt dans de nombreuses applications environnementales et industrielles. Bien que ce problème soit étudié depuis de nombreuses années, sa compréhension n’est pas encore suffisante pour la création de modèles de transfert de masse réalistes (de type RANS ou LES sous maille), en particulier en présence d’une phase liquide à rhéologie complexe. Ce travail expérimental a pour but l’étude des aspects fondamentaux du transfert de masse turbulent à une interface plane horizontale entre du dioxyde de carbone gazeux et une phase liquide newtonienne ou non, agitée par une turbulence homogène quasi isotrope. Les milieux liquides non newtoniens étudiés sont des solutions aqueuses d’un polymère dilué à des concentrations variables et aux propriétés viscoélastiques et rhéofluidifiantes. Deux méthodes de mesure optiques permettant l’obtention du champ de vitesse de la phase liquide (SPIV) et de concentration du gaz dissout (I-PLIF) sont couplées tout en maintenant une haute résolution spatiale, afin de déduire les statistiques de vitesse et de concentration couplées dans les premiers millimètres sous la surface. Une nouvelle version de I-PLIF est développée pour les mesures en proche surface. Elle peut également s’appliquer dans différentes études de transfert de masse. La turbulence de fond est générée par un dispositif de grille oscillante. Les mécanismes de production et les caractéristiques de la turbulence sont étudiés. L’importance de la composante oscillante de la turbulence est discutée, et un phénomène d’amplification de l’écoulement moyen est mis en évidence. Les mécanismes du transfert de masse turbulent à l’interface sont finalement observés pour l’eau et une solution de polymère dilué à faible concentration. L’analyse conditionnelle des flux de masse turbulent permet de mettre en évidence les évènements contribuant au transfert de masse et de discuter de leur impact relatif sur le transfert total. / The study of turbulence induced mass transfer at the interface between a gas and a liquid is of great interest in many environmental phenomena and industrial processes. Even though this issue has already been studied for several decades, its understanding is still not good enough to create realistic models (RANS or sub-grid LES), especially when considering a liquid phase with a complex rheology. This experimental work aims at studying fundamental aspects of turbulent mass transfer at a flat interface between carbon dioxide and a Newtonian or non-Newtonian liquid, stirred by homogeneous and quasi isotropic turbulence. Non-Newtonian fluids studied are aqueous solutions of a model polymer, Xanthan gum (XG), at various concentrations, showing viscoelastic and shear-thinning properties. Optical techniques for the acquisition of the liquid phase velocity field (Stereoscopic Particle Image Velocimetry, SPIV) and dissolved gas concentration field (Inhibited Planar Laser Induced Fluorescence, I-PLIF) are for the first time coupled, keeping a high spatial resolution, to access velocity and concentration statistics in the first few millimetres under the interface. A new version of I-PLIF is developed. It is designed to be more efficient for near surface measurements, but its use can be generalized to other single or multiphase mass transfer situations. Bottom shear turbulence in the liquid phase is generated by an oscillating grid apparatus. The mechanisms of turbulence production and the characteristics of oscillating grid turbulence (OGT) are studied. The importance of the oscillatory component of turbulence is discussed. A mean flow enhancement effect upon polymer addition is evidenced. The mechanisms of turbulent mass transfer at a flat interface are finally observed in water and low concentration polymer solutions. A conditional analysis of turbulent mass fluxes allows to distinguish the type of events contributing to mass transfer and discuss their respective impact in water and polymer solutions.

La sûreté des réacteurs nucléaires constitue une condition primordiale et nécessaire à leur industrialisation. Pour les réacteurs à neutrons rapides refroidis au sodium, cette sûreté passe par la possibilité de détection d’une fuite d’eau dans le sodium au niveau du générateur de vapeur à des stades très précoces de leur déclenchement. Le présent travail de thèse s’inscrit dans le cadre du développement d’une technique non intrusive pour la détection d’une réaction sodium-eau dans le générateur de vapeur d’un réacteur nucléaire refroidi au sodium. On désire identifier le bruit vibratoire de cette réaction à partir de mesures d’accélération sur la virole externe du générateur. Cependant, les vibrations dues à la fuite peuvent être noyées dans le bruit généré par l’écoulement du sodium ou par d’autres sources d’excitation. Afin d’augmenter, le rapport signal à bruit, on s’intéresse à l’apport d’un filtrage spatial par formation de voies pour ce type de système. Ce dernier se caractérise notamment par la présence d’un couplage structure-fluide lourd (c.-à-d. virole-sodium) et des modes vibroacoustiques qui apparaissent aux mêmes fréquences d’intérêt pour la détection de la fuite. Pour réaliser cette étude, nous considérons une maquette expérimentale composée d’une conduite cylindrique reliée à un circuit hydraulique par deux brides très rigides. La source à identifier est simulée par un hydrophone de taille réduite en mode émetteur positionné à l’intérieur de la conduite par un dispositif mécanique dédié à cet effet. Le bruit perturbateur est induit par l’écoulement de l’eau, supposé turbulent à partir d’un certain débit. L’antenne est composée de vingt-cinq accéléromètres placés sur la conduite d’essai. L’objectif consiste à traiter simultanément les signaux accéléromètriques pour faire ressortir le signal de la fuite tout en rejetant le bruit dû à l’écoulement. Deux types de traitement par formation de voies sont considérés : la formation de voies conventionnelle (dites de Bartlett) et une formation de voies optimisée qui vise à maximiser le rapport signal à bruit de la fuite. / The safety of nuclear reactors represents a necessary and vital condition for the exploitation of nuclear plants with liquid-cooled cores. This safety passes by the ability to detect and anticipate the earliest stages of a water leak into sodium within the heat generator. The study detailed in this manuscript has been initiated in a framework aiming to develop nonintrusive detection techniques for sodium water chemical reactions. Its main goal is to uncover the signal of a water leak into sodium, based on the measurements of the vibratory field recorded by the means of accelerometers externally mounted on the heat generator shell. However, such a spectrum is masked by the plant general background noise, especially that generated by heavy fluid flow (i.e. sodium flow) during actual operating times of the power plant. In order to increase the signal-to-noise ratio of the leak, beamforming technique for the acceleration measurements of the mechanical system is considered. The aforementioned system is characterized by a strong non-linear coupling between the the heat generator’s cylindrical shell and the heavy fluid in motion. In particular, fluid motion and the acoustic emission of the leak seem to excite some eigen modes of the cylinder, at these same frequencies where the acoustic signature of the leak is at its highest amplitude. For the purpose of our study, a cylindrical mock-up connected by some very rigid links to a hydraulic circuit is considered. A hydrophone emission excites the mock-up from within and is being accounted for the acoustic leak. Water flows inside the cylinder at turbulent Reynolds number. An array of twenty-five accelerometers is mounted on the mock-up shell using ceramic insulators. The main goal is to combine the twenty-five signals in such a way that allows the increase of the SNR for the acoustic source while rejecting water flow noise. Two beamforming techniques are applied and compared: classical Bartlett beamforming as well as optimized beamforming for SNR maximization (Max SN