FLUID MECHANICS

Author: CERQUEIRO GARCÍA DANIEL.
Year: 2003.
University: VIGO [www.uvigo.es].
Place of defense: FACULTAD DE CIENCIAS.
Place of preparation: FACULTAD DE FISICA UNIVERSIDAD DE SANTIAGO.

Summary: This work aims to end collaborate in the establishment of a local model of generation (in this case the SWAN), which stir corretamente silver mainland form. This has had a landmark case of 4 months with the boundary conditions obtained from WAM and forcing wind model atmospheric ARPS. The results obtained by the application of SWAN, from a general point of view, are better than those of WAM significant height (HS) and average duration (TM), while in period of peak (TP) of WAM are slightly better . Tests have also been conducted with different parametrizations físics that the model presented for different terms SWAN strong.

Author: RIBAS PRATS FRANCESCA.
Year: 2003.
University: POLITÉCNICA DE CATALUÑA [www.upc.edu].
Place of defense: FISICA APLICADA.
Place of preparation: FISICA APLICADA.

Author: BARDERA MORA RAFAEL.
Year: 2004.
University: COMPLUTENSE DE MADRID [www.ucm.es].
Place of defense: FACULTAD DE CIENCIAS FÍSICAS.

Summary: This paper presents a new method for determining the coefficient of drag, from the measure of speed on the trail, using the technique of Velocimetría Particle Image, known as IVP. The novelty of the method lies in precisely determining the coefficient of drag using the PIV technique, which is commonly used to measure the speed local flujos.La advantage of the proposed method compared with conventional techniques is that PIV is not intrusive and simultaneously measure also allows the distribution of velocities in the wake of a plane. Experiments conducted with several cylinders and plates have helped prove the method proposed as suitable for measuring the coefficient of resistance, because the results are consistent with those outlined in the literature obtained by other methods for the same experiments.

Author: CABEZAS MARTÍN M. GUADALUPE.
Year: 2004.
University: EXTREMADURA [www.unex.es].
Place of defense: ESCUELA DE INGENIERÍAS INDUSTRIALES DE BADAJOZ.
Place of preparation: ESCUELA DE INGENIERÍAS INDUSTRIALES.

Summary: The methods of measuring the surface tension get the most widely used value of this magnitude analyzing digital images of drops and bubbles axisimétricas. The value of the surface tension is calculated by adjusting the theoretical predictions of how to drop the experimental contour shown in the image. The pilot contour extracted from the image using techniques for detecting contours standards. These measurement methods presented certain limitations that sometimes lead to erroneous results. On the one hand, if the images are of poor quality or have a poor contrast, the boundary drawn by the detection techniques is wrong. On the other hand, even if the pictures are good quality, the shape of the drop may not be sufficiently sensitive to variations of the magnitude that we want to measure. The purpose of this dissertation is to develop a method of measuring the surface tension that has not provided such limitations. First, it developed a technique for detecting contours guided to replace the currently used. This technique uses the theoretical prediction of contour in the search for the pilot within contours of the image. In theory the validity of this technique is found on experimental images of different qualities, comparing the results with those provided by current techniques. Furthermore, the thesis explores the use of a fluid configuration, called liquid bridge, as alterantiva to drop axisimétrica. We have studied the sensitivity of this configuration smooth face changes in the value of the surface tension, comparing the results with those obtained for a drop axisimétrica. Finally, a series of experiments are performed with liquid bridges, and compares the values of the surface tension measured using current techniques and developed in this thesis.

Author: ALAMO DE PEDRO JUAN CARLOS DEL.
Year: 2004.
University: POLITÉCNICA DE MADRID [www.upm.es].
Place of defense: ESCUELA TECNICA SUPERIOR DE INGENIEROS AERONAUTICOS.
Place of preparation: ESCUELA TECNICA SUPERIOR DE IGENIEROS AERONAUTICOS.

Summary: This thesis investigated the origin and organizational structures that are in large regions logarithmic and external turbulence parietal. These structures contain most of the turbulent kinetic energy, are responsible for a significant fraction of the resistance and friction play a key role in the turbulent transport. Despite its importance, there is much less information about the large structures on small structures in the region of the wall. This is due to the difficulty of analysis. For our study we use the results of new experiments numerical turbulent flow in channels drawings made Reynolds numbers, plus high and computational domain larger than the existing ones so far. Some of the results of these experiments is available on the site http://torroja.dmt.upm.es/ftp/channels. Housed in Spain or in the mirror http://davinci.tam.uiuc.edu/data/channels. Housed in America. UU. It is also possible to access the entire database by following the procedures described in htlp: / / davinci.tam.uiuc.edu / content / Channeldata.html. The results indicate that the turbulent fluctuations of speed in the directions parallel to the wall are arranged to form large structures, whose size scale with the thickness of the flow. These structures, which we called modes global penetrate from the center channel to the area of the wall, which explains the scaling incomplete spectrum of energy in this region. We have shown that the scale of speed modes is the global average speed in the center of the channel, reaching a joint escalated in the region outside, it boils down to - to. We have studied the properties of similar spectrum of energy in the regióN intermediate flux, finding anomalies in the relationship between the lengths yanchuras of large scales. The shape of anomalous scaling-suggests that the structures are of great trails left by compact objects of vyW, to be broadcast under the influence of the turbulent viscosity -. This leads to failure of similarity logarithmic corrections in several ranges of the spectrum classic, which have been tested using data from our simulations and laboratory experiments from different streams wall at Reynolds numbers greater than ours. It has analyzed the origin of the large structures of the layer logarithmic, finding that this is in intensive ejections from the region of the wall. For this analysis we have developed a method that allows isolate and classify the ejections, based on the evidence that are associated with clusters of eddies embedded in the wall that reaches the intermediate region. The velocity field conditioned medium to each of these ejections structure reveals a very long, whose geometry is consistent with the anomalous scaling spectrum in the intermediate region. These results provide a basis for our structural model based on stelae.

Author: GAMEZ MONTERO PEDRO JAVIER.
Year: 2004.
University: POLITÉCNICA DE CATALUÑA [www.upc.edu].
Place of defense: ETSEIT. CAMPUS TERRASSA.
Place of preparation: ETSEIB, Campus SUD.

Summary: In recent years, research are highly motivated by anything that helps to reduce the noise emitted by the systems oleohidráulicos. As has been demonstrated by various researchers, the origin of these disturbances is to be found in the process of generating the beats flow as a feature inherent own actual operation of the pump, and therefore, any progress in this area must inevitably pass by the simulation of the dynamic behavior of these units. This research work is confined to the study of a bomb gerotor, in order to have a comprehensive study, which includes coherent and unified geometry profiles trocoidales gear, the kinematics of gear, features volumetric, tensions in the contact points and their dynamic behavior. The geometry and kinematics of gear gear trocoidal are drawn analytically by equations of the profiles of teeth and the line of contact points on the basis of four key parameters, laying the basis for the development of subsequent studies. The volumetric characteristics have been developed, as a point and compared using two methods: one that is based on numerical analysis integral-derivativo, and the other was based on an analysis derivativo-integral in its two possible versions, analytical and numerical. The similarity of results obtained using these methods is very satisfactory. In calculating tensions contact, have also been proposed and developed two methods of calculation. A first method part of the theory developed by Colbourne, implemented with an option for the calculation of the contact points. This method is suitable for external gears where the profile of the lobe or tooth is adjustable to a circular arc. Since its implementation is very restrictive, it was considered appropriate to propose an alternative method implemented with the method of Finite Elements much more generic. The latter has provided the values of the magnitude of the maximum voltage contact more reliable and has demonstrated a high level of flexibility that greatly facilitates the process of optimizing the geometrical shape of the profiles of teeth. The behavior fluidodinámico the pump gerotor is proposing a method of studying physics simulation (Method BondGraph) that falls directly in the calculation of the flow pulsations. The simulation model fluidodinámico presents a mathematical analysis that part of submodelo a camera and complete interaction engranaje-bomba and leakage flow. This methodology was developed numerical technique contrasts with a pilot (Method of âFuente Secundariaâ) that allows direct comparison of the results obtained by both tracks. There are two test beds used and developed in accordance with ISO 10767, one in the Fluid Power Center at the University of Bath, and another designed and built specifically for this thesis in the laboratory from the center LABSON. It is noted that the amplitude (peak to peak) of the beats flow obtained by the pilot in the Fluid Power Center are lower and more regular than those achieved in the laboratory from the center LABSON. The comparative results of the pulse flow âideal' obtained from the simulation of flow and pulsation of the experiments appear to be very precise in its way, despite differences in amplitude. The simulation model shows that describes with great accuracy interaction engranaje-bomba and leakage flow pump gerotor

Author: Valencia Merizalde Leonardo.
Year: 2004.
University: ROVIRA I VIRGILI [www.urv.cat].
Place of defense: E.T.S. d'Enginyeria Química.
Place of preparation: E.T.S. d'Enginyeria Química.

Summary: This paper explores the convection Rayleigh-Bénard in cubic cavities without leaning on the horizontal and heated below for Rayleigh numbers both within the state and within the stationary turbulent regime. Initially explores the impact of changes in physical properties with temperature on the structure of flow, the mechanism of transport and heat transfer for two low Rayleigh numbers where the flow is laminar and steady and for a Rayleigh in the turbulent regime. For this study numeric water was used as convective fluid and it was assumed that the side walls of the cavity were perfectly conductive. Subsequently have been identified numerically structures average flow time for five Rayleigh numbers within the turbulent regime. Due to the similarity of four such structures only two of them were verified experimentally. For the calculations are assumed approximate Boussinesq due to temperature differences in testing were sufficiently low as to consider constant physical properties according to the results obtained in the above analysis. In order to reproduce the most experimental conditions, the numerical results were obtained taking into account the thermal conductivity of glass in the sidewalls. The visualization of the flow structures and the measurement of velocity fields in the vertical plane middle of the cavity was performed using the method IVP. These results enabled us to validate the results with simulations, comparing both topologies flow velocity values as timely profiles within the plane analyzed. Of the numerical results found that even with rates varying physical properties of the fluid between the hot and cold walls far above the standard normally used in both laminar and turbulent regime, the flow structures and conditions of heat transport, they are not significantly affected by this change. Moreover structures and temporary average flow velocity values obtained are consistent numerically significantly with the corresponding experimental measures for Rayleigh numbers analyzed.

Author: HAMDY DOWEIDAR DOWEIDAR TAKY EL DIN MOHAMED.
Year: 2004.
University: ZARAGOZA [www.unizar.es].
Place of defense: CENTRO POLITÉCNICO SUPERIOR.
Place of preparation: CENTRO POLITÉCNICO SUPERIOR.

Summary: So far the finite element simulation of flows with terms source and estimating a-posteriori error in computational fluid dynamics has been a challenge. Through this thesis make progress on three main themes. The first presents a variational formulation that combines two types of integrated stabilizing called SMS / GSGS. The two free parameters are chosen to get superconvergencia dimensional. When the method is applied to problems of multi-dimensional flow, the local stability of this method is far better than current methods stabilized. In addition, the limit advectivo-difusivo and functional spaces for Linear pieces, the method retrieves the classic method of SUPG. The second issue is the analysis of the most common methods transient Galerkin, SUPG and SGS-dimensional equation for the advección-difusión-reacción. The same analysis also applies to the method SGS / GSGS. This analysis includes methods stabilized finite element semi-discretos and discontinuous in time and space, as well as several versions of methods predictor multi-corrector. It proposes and discusses a new treatment of the terms sources. The technique consists in treating diagonally implied terms sources negative or disipantes to increase the stability and accuracy of the explicit methods. In addition, it will be shown that the accumulation of mass can improve the accuracy of methods explicit. In the last issue investigates a new estimator explicit a-posteriori error that is deducted from the variational theory of multiscalas. The estimator of error is based on an approximation of the Green's function which reproduces the error according to the dual problem. The technique applies to the transportation of fluid flow, showing that flows dominated by convection, the rate of effectiveness is independent of the diffusion coefficient. It finds that the estimator multiscala of error is exactly where these scales are used for the one-dimensional nodalmente exact solutions with space finite element linear pieces, both for the case of waste constant linear pieces as pieces. One of the features of this technique is that it gives the scales appropriate for the estimation of a-posteriori error in any standard interest, as L_2, H ^ 1, energy and the rule L_infty.

Author: JIMÉNEZ TORRECILLA SANTIAGO.
Year: 2004.
University: ZARAGOZA [www.unizar.es].
Place of defense: CENTRO POLITECNICO SUPERIOR.
Place of preparation: LITEC / CENTRO POLITÉCNICO SUPERIIOR.

Summary: Particulate emissions from biomass combustion have been experimentally characterized in a laboratory facility, for a variety of fuels and combustion conditions. Two particle modes were found in all the cases; while coarse particles essentially retained the original fuel ash composition, those in the submicron mode were composed only of alkali (mainly potassium) sulphates and chlorides in different proportions. A new probe (AQPS), based on aerodynamic-quenching, and a TEM allowed for the detailed study of the fine particle formation mechanisms. Evidence has been found for the existence of "formation steps": K2SO4 nucleates first in the flue gas cooling process, whereas KCl condenses on these nuclei at lower temperatures. This is in good agreement with simple thermodynamic equilibrium calculations and with theoretical models by other authors. Experimental evidence is shown, also, for the kinetic limitation of SO2 oxidation in the route to sulphate formation. The final chloride/sulphate ratio in fine particles is greatly affected by variations in the SO2 - O2 concentrations in post-combustion gases. Moderate values of extra SO2 gas for usual O2 concentrations, as those attainable in co-firing systems, result in the total disappearance of chlorine from these particles, a fact that has been experimentally confirmed in co-combustion tests in this study. This gives more basic support to field results on deposit characteristics and boiler corrosion previously reported, and opens a way for minimizing the chlorine-associated boiler operating problems in biomass combustion. The effectiveness of three different particle sampling methods was comparatively studied in post-combustion flows containing significant amounts of inorganic condensable vapours. While the use of the AQPS probe resulted in all the cases tested in the formation of ultrafine particles by rapid vapour nucleation, the artifacts caused by the probes based on N2-dilution and thermophoresis presented great variations, thus making the interpretation of the results obtained difficult.

Author: RODRÍGUEZ PÉREZ DANIEL.
Year: 2005.
University: NACIONAL DE EDUCACIÓN A DISTANCIA [www.uned.es].
Place of defense: FACULTAD DE CIENCIAS.
Place of preparation: UNED.

Summary: The deposits of particles whose transport is described by the equation of convective diffusion, presented a density varies, depending on which of these two mechanisms is more important. The relative importance is described by the number of dimensionless Péclet (Pe). In the case purely diffusive (Pe = 0), deposits are fractals, and their density decreases with distance from the surface of deposits. These deposits are also muyn graft and its roughness increases over time according to a power law. In the case ballistic (Pe = infinity), the density of the deposits does not vary with height. However, the roughness of the interface grows with time also according to a power law. The objective of this thesis is to describe as changing the structure of deposits from the branched fractal until the compact, as increasing numbers of Péclet. First, it developed two models prbabilisticos motion, a tight mesh cubic compatible with the equation convección-difusión, according to the statistical interpretation of it. One of them, based on the concept of probability of transition between cells is limited to low numbers Péclet (Pe increased 4). The other ees the general validity, and the algorithm separates each step in a temporary stage and convecting another difusiva. Secondly, deposits are generated and simulated particle is characterized his profile densities, as measured by the average density depending on the number of Pélet. This unit is interpreted in terms of a model fractalidad local introducing a characteristic scale of the order of 1 + A / Pe (A is a parameter of order unity). This interpretation is supported by the collapse of the curves of counting boxes and pot delas curves profile densities, the referirlas at that level. Finally, the interface is characterized growth of deposits, that is the active area or region in which they are deposited particles. The lengths characteristics (anchua and longitude correlation horizontal) are calculated by weighting with the likelihood of arrival of new particles to the sites of active growth. These lengths evolve according to laws powers of the time: the exponents relate to the universality KPZ. For different númeos de Péclet obtained exponenetes numéricamentes similar. Thanks to this or niversalidad behavioral curves evolution of correlations can be horizontal and vertical collapse rescalando time with density, under the Péclet, and with a width depending on the scale characteristic of the interior of the tank. The main contributions of this thesis are: proposed two different models to simulate the movement balístico-difusivo particles adhered to a mesh cubic introduces a model fractalidad local to describe the interior of these deposits formed by particles, and has The results of these numerical simulations can be applied to the procurement of materials required in volume properties (density, porosity) and area (roughness, active area) through dynamic control of the deposition of the constituent particles.

Author: ALBETS CHICO XAVIER.
Year: 2005.
University: POLITÉCNICA DE CATALUÑA [www.upc.edu].
Place of defense: ETSEIB-UPC.
Place of preparation: ETSEIB, EDIFICI H PLANTA 8 Campus SUD.

Summary: Turbulent flows are extremely usual in both natural and industrial environments. Hence, their analysis and resolution are of great interest for the development and improvement of industrial processes and products; for the comprehension of atmospheric, oceanic and environmental phenomena; for the development of new energy sources (or for the optimization of the traditional ones); as well as for the advance of living conditions, among a lot of diverse aspects. Unfortunately, under real cases, complexity regarding turbulence makes impossible its mathematical resolution and implies a highly costive computational resolution (CPU time-consumption, memory storage, etc.). Because of all these aspects, basic research on turbulence models is needed in order to convert the deeply interesting turbulent cases into tractable situations from engineering and computational points of view. Present thesis is based on analysis of time-averaged mathematical models or âReynolds-Averaged Navier-Stokes❠(RANS) turbulence models. More concretely, this work is built on the use of Eddy-viscosity two-equation models. These mathematical techniques have became, up to present moment, the most usual methods for turbulence resolution in the surrounding of the engineering field, because they provide a compromise between accuracy and tractability. From a numerical point of view, computational discretizations have been developed over the well-known finite volume techniques in combination with the SIMPLE-like methodology for the resolution of the mass-momentum coupling. Such techniques are the most usual methods for both Computational Fluid Dynamics (CFD) and turbulence modelling by the international scientific community. Main purposes of the work are spread, on the one hand, throughout the acquisition of a know-how regarding Eddy-viscosity two-equation models under basic forced and natural convective cases and, on the other, over the numerical and computational analysis of the application of such models for resolution of orthogonal and non-orthogonal geometries. Additionally, the achievement of these goals has involved mathematical and numerical contributions to the general purpose in-house code developed by Heat and Mass Transfer Technological Center (CTTC) where this thesis has been carried out. Present work emphasises the effort devoted to research on abilities and drawbacks of this kind of models for predict benchmark turbulent cases (specially regarding Wall-Functions (WF) methods for forced convection cases and regarding Low Reynolds-Number (LRN) models for natural convection situations). As a result, the thesis has provided detailed considerations about possibilities of such models for prediction of exposed situations. Diverse problems have appearedand different suggestions (from both numerical and physical points of view) have been presented. New ideas with reference to Wall-Function treatments in combination with two-equation models as well as drawbacks and points regarding code generalization over non-orthogonal geometries have been given. As conclusion remark, it has to be said that present work has achieved the previously proposed goals, throughout effecting a rigorous work regarding both numerical (or computational) and mathematical aspects. By this way, an appreciable know-how related to treated models under diverse turbulent phenomena has been obtained.

Author: OROZCO PIÑEIRO CRISTIAN.
Year: 2005.
University: POLITÉCNICA DE CATALUÑA [www.upc.edu].
Place of defense: ESTSEIB.
Place of preparation: ETSEIB, EDIFICI H PLANTA 8 Campus SUD.

Summary: Problems where complex geometries or where moving boundaries exist arises in numerous scientific and engineering applications. Illustrative examples can be found in the thermal machinery field, such in reciprocating compressors, in internal combustion engines, or in air handling units. In the design and optimisation studies of theses systems, the detailed analysis of the heat and mass transfer phenomena becomes an essential issue. From the Computational Fluid Dynamics (CFD) point of view, a common characteristic to all these problems is the need to develop specific treatments from the standard CFD methodologies to deal with the geometry of the domain under study. The works conducted in this thesis tackle the modelization and the numerical methodology required for the fluid flow and heat transfer computation within this range of applications. The flow computation in domains with moving boundaries is approached from the moving grid method. In this method, the governing equations are discretized over control volumes which change their geometry at every time-step. This non-Eulerian, or arbitrary Lagrangian-Eulerian (ALE), formulation introduces the necessity to satisfy the Space Conservation Law (SCL) together with the rest of governing equations. Failing in preserving this principle might lead to additional sources of error in the solutions, and, as it is demonstrated in this work, to introduce oscillatory behaviour in the convergence of the error. The studies conducted in this thesis are focused on the implementation of the moving grid method on Finite Volume (FV) computations, and on the analysis of the effects that parameters such as the mesh velocity, or the time and the SCL discretization, have on the accuracy and on the asymptotic or oscillatory convergence of both the space and time discretization errors. Test cases analysed include manufactured solutions, and the incompressible and compressible flow on different piston-cylinder configurations. The computation of fluid flow and heat transfer in complex geometries is approached from the porous media modelization. First, fundamental representative geometries of the complex element are identified and geometrically characterised. Then, their effects on the fluid flow are modelled and introduced in terms of a distributed resistance to the fluid flow and a source/sink of temperature, humidity, turbulent kinetic energy, etc. This method, usually referred as the Porosity and Distributed Resistances method (PDR), is approached in this work from both the common quasi-continuum approximation, and from an alternative discrete formulation. The second approach incorporates additional empirical information in the model. The final aim is, by adding additional empirical information, to provide more accurate flow estimations in those cases where such information is available. The studies conducted in this work have been planned as a preliminary study on the mathematical and numerical aspects of Finite Volume (FV) implementations of the PDR approach, based on both the quasi-continuum and the discrete formulations. Special attention has been paid to the treatment of discontinuities through the free fluid-porous media interface. Finally, a common aspect, not only to CFD calculations, but also to many calculations obtained by numerical means, is the assessment of the credibility of the results obtained. The work developed is focused on the verification of numerical solutions, this is, in the development and the assessment of reliable tools for the quantification of the numerical uncertainties on computational solutions. Attention has been focused on the study of the suitability of systematic grid refinement methods, based on the Richardson extrapolations family 8 and the 349 Grid Convergence Index (GCI), to provide accurate and conservative estimations of the error band on both steady and transient solutions which reports convergence far from the asymptotic range.

Author: SANTIAGO DEL RIO JOSE LUIS.
Year: 2005.
University: POLITÉCNICA DE MADRID [www.upm.es].
Place of defense: ESCUELA TÉCNICA SUPERIOR DE INGENIEROS AGRONOMOS.
Place of preparation: ESCUELA TÉCNICA SUPERIOR DE INGENIEROS AERONÁUTICOS.

Summary: Air pollution in urban settings has become in recent years one of the most important environmental issues. For this reason, both the authorities and the scientific community have paid special attention to the study of the air quality inside the cities. One example is the "Air pollution at street levels in European cities" (2006) conducted by the European Environment Agency (European Environment Agency), which analyzes air quality in streets of 20 European cities. The highest peaks of concentration appear inside the cities, especially in the interior of the narrow streets, and traffic source of emission of pollutants. The main objective of this study is to contribute to the understanding of the physical processes (circulations of air and dispersion of pollutants) which occur in the interior of the streets and the urban boundary layer. To this end, model simulations are performed on CFD geometries similar to a city. The models are based on equations RANS, with different models of turbulence type k-epsilon. It also makes use of experimental studies previously conducted to validate the model results. Thus, in Chapter 3 discusses the circulations of air into symmetrical streets and symmetrical 2D. This demonstrates the influence of tallest building on the flow of air into the streets. In Chapter 4 discusses the flow of air over barrier belts with different porosity. Thus discusses barriers belts as a means of mitigating dust emissions in areas acopios material pulverulento. It provides for the porosity of the barrier valued at 0.35 as it produces an optimal shielding, and therefore, a good reduce emissions of dust. In addition, this study also allows contrasting schemes closing turbulent as it has a good database experimental wind tunnel (IDR). Continuing with the scenarios in 2D, Chapter 5 presents a new model largragiano particle (SLP-2D) developed in the thesis. To check their results were two types of validation: 1-Pretend the concentration of pollutants in an experiment in wind tunnel (Meroney et al., 1996). 2-Pretend concentration in two real streets. To represent the city in a manner closer to reality, in chapters 6 and 7 will simulate the flow of wind and the dispersion of pollutants on an array of cubes representing a city. Chapter 6 focuses on the characteristics of the flow pattern and validation measures wind tunnel, while Chapter 7 is focused on the dispersion of pollutants looking relations between the position of the emission and concentration of the tracer. Finally, in Chapter 8 describes major findings and future lines of inquiry.

Author: TRIAS MIQUEL FRANCESC XAVIER.
Year: 2006.
University: POLITÉCNICA DE CATALUÑA [www.upc.edu].
Place of defense: SALA de conferències etseiat.
Place of preparation: ETSEIB, Edifici H LAB. TERMODINAMICA, PLANTA 8 SD.

Summary: The main purpose of this thesis has been to contribute to the development of numerical algorithms for the Direct Numerical Simulation (DNS) and regularization modelling of turbulent flows on loosely coupled parallel computers. In the last decades, DNS has become a very important area of contemporary fluid dynamics, because it is an essential tool to give new insights into the physics of turbulence and to provide indispensable data for future progresses on turbulence modelling. In the first chapter, a general symmetry-preserving discretization method is presented. The formulation presented intends to lead to a generalisation of the work of Verstappen and Veldman (J.Comput.Physics 187 (2003) 343) for general unstructured meshes. The operator formulation used here is based on the 'shift' transformation idea recently proposed by Kicken et al. (J.Comput.Physics 208 (2005) 704). It allows matrix operators for collocated variables to be easily transformed into matrix operators for staggered variables while preserving symmetries. The basic idea behind remains the same: mimicking the crucial symmetry properties of the underlying differential operators, i.e., the convective operator is approximated by a skew-symmetric matrix and the diffusive operator by a symmetric, positive-definite matrix. In chapter 2, a parallel Direct Schur-Fourier Decomposition (DSFD) algorithm for the direct solution of arbitrary order discrete Poisson equations on parallel computers is presented. It is based on a combination of a Direct Schur method and a Fourier Decomposition and allows to solve each Poisson equation to machine accuracy using only one global communication episode. Thus, it is perfectly well suited for loosely coupled parallel computers, that have a high network latency compared with the CPU performance. Several numerical examples illustrating its robustness and scalability on a PC cluster with a conventional 100 Mbits/s network are presented. Numerical methods presented in the first two chapters are used in chapter 3 to perform direct numerical simulations. The symmetry-preserving discretization described in chapter 1 is specialized for orthogonal Cartesian grids and the DSFD algorithm is used to solve efficiently the Poisson equation. In chapter 3, a set of complete 2D and 3D direct numerical simulations (DNS) in a differentially heated air-filled cavity (DHC) of aspect ratio 4 with adiabatic horizontal walls is presented. Although the physical phenomena is 3D, due to its prohibitive computational costs the majority of the previous DNS of turbulent and transition natural convection flows in enclosed cavities assumed a 2D behaviour. An overview of the numerical algorithm and the methodology used to verify the code and the simulations is presented. As expected, significant differences are observed between 2D and 3D results. For 2D simulations the oscillations at the downstream part of the vertical boundary layer are clearly stronger, ejecting large eddies to the cavity core. In the 3D simulations these large eddies do not persist and their energy is rapidly passed down to smaller scales. In the last chapter, the new DNS results obtained in the chapter 3 are used as reference solution for the development of better turbulence models. We consider a new class of regularization (smooth approximations) of the nonlinear (convective) terms. The regularization method basically alters the convective terms to reduce the production of small scales of motion by means of vortex stretching. In doing so, we propose to preserve exactly the symmetry and conservation properties of the convective terms. This requirement yields a novel class of regularizations that restrain the convective production of smaller and smaller 8 scales 374 of motion by means of vortex-stretching in an unconditionally stable manner. The regularization model is successfully tested for the 3D DHC (Ra=1e10, Pr=0.71 and height aspect ratio 4) whose DNS results have been presented in chapter 3.

Author: SALAZAR MARIN EDGAR ALONSO.
Year: 2006.
University: POLITÉCNICA DE CATALUÑA [www.upc.edu].
Place of defense: Sala de Juntes de l'ETSEIT.
Place of preparation: ETSEIB, PAVELLÓ D Campus SUD.

Summary: This thesis is related to a macro project whose title is 'for New design and manufacturing processes for high pressure fluid power products' -2004-2008 (acronym: PROHIPP), which is co financed by the European Union and where take part 22 organizations of 8 different countries between small and medium European companies, research centres and universities. The main objective of the thesis is to study the load capacity of the hydraulic cylinders, determining the factors that influence of important way over it. Critical analysis of some publications in relation to the subject of load capacity of cylinders were made, giving importance at which could offer greater contributions for the project. Norm ISO 13725 (Hydraulic fluid power -cylinders- method for determining the buckling load), presents a difficult interpretation of the raised formulations, and its proposed model omits as important factors as the initial imperfection. The project develops an algorithm that allows an easy application of this Norm using any boundary conditions. The project proposes a model that describes the behaviour of the pined-end hydraulic cylinders under axial load, involving present factors in all real application like: initial imperfection (misalignment between piston rod and tube-cylinder), own weight, moments by friction and/or eccentricity in the end points of the actuator. The model allows mainly determining the load capacity of a hydraulic cylinder, calculating the maximum permissible load, which generates a maximum stress throughout the actuator, equivalent to the yield stress of the material. In addition, the model allows determining curves of deflection, moments and stresses throughout the hydraulic cylinder under a given load. A test bench has been designed and constructed to validation of the obtained theoretical results with the proposed model. This bench allowed the creation of an experimental data base of more than 100 destructive tests for the determination of the permissible maximum load of cylinders. The bench has allowed the measurement of important variables in the phenomenon, like deflection of the hydraulic cylinder, load applied by means of the internal pressure of the fluid into the tube-cylinder, displacement of the piston rod within the tube (compressibility of the oil), stresses in the piston rod (microdeformations), etc. A etestf cylinder has been used as base of experimentation and validation of the proposed model. As result of the theoretical-experimental analysis of the project has been able to conclude the following: The inclusion of the own weight (100 N) of the etestf cylinder has meant only a reduction of 2% of its load capacity. The variation of the load capacity due to the initial imperfection (due to the tolerances of a good design and a good manufacture) is much smaller than the variation of load capacity due to the real boundary conditions that are presents in the pin-ends. The wear of the guide ring in a 5% of its initial thickness (effect that takes place basically in the first 1000 cycles) represents a decreasing of the load capacity of a 10%. It is necessary to mention that from the 1000 cycles there are no significant variations in the wear of the guide rings. As final conclusion and recommendation of the project can be affirmed that the friction effects always must be diminished to the maximum, as much if it is favourable or against the load capacity. In those applications where the hydraulic cylinder drives a mechanism and the effects of the friction go against the load capacity, is advised to calculate the hydraulic cylinder with a external applied moment in the pin-end of the side piston rod (weaker side) and of equal value to the product of the applied axial load multiplied by an apparent eccentricity, of value 2 m dP (ÆÃ: friction coefficient, dP diameter of the pin).

Author: JIMENEZ GARCIA MIGUEL ANGEL.
Year: 2006.
University: POLITÉCNICA DE MADRID [www.upm.es].
Place of defense: E.T.S. ING. INDUSTRIALES.
Place of preparation: E.T.S. ING. INDUSTRIALES.

Summary: Accidents in Unit 2 of the core of the Three Mile Island in the United States (March 28, 1979) and the Unit 4 of the Chernobyl plant in Ukraine (April 26, 1986) prompted research Severe accidents involving damage to the nucleus, and in particular the possibility of failure of the last barrier for containment effects of a hydrogen explosion. It led research programs to analyze the behavior of hydrogen and its potential control. In industry, there are a wide variety of systems based on the catalytic reaction H2/O2, as recombinadores autocatalíticos liabilities marketed by various manufacturers and installed in existing nuclear power stations in various countries for the control of hydrogen. Its main feature is the operating liabilities, or their ability to recombine them effectively hydrogen by passing its flow through a natural convection driven by the heat released in a chemical reaction catalytic automantenida on its surface (mixed reaction). On the other hand, the risk of spread in the gas phase should be seen as a possible limitation of these devices to high concentrations of hydrogen. In this work doctoral thesis has developed a detailed model to replicate the processes of recombination of H2 and CO on PARs of parallel plates, through the coupling of surface chemistry and transport of mass and heat between mixtures H2/CO/aire / vapor/CO2 and vertical plates coated with Pt. This model PARUPM is based on a simplified outline of Deutschmann for combustion of methane and surface analysis Elenbaas for the heat transfer between parallel plates induced flotation. The model allows isolated analyze the startup phase of the RAP until it reaches its steady operating conditions fixed in the outer atmosphere, and has been contrasted against the main predictive models based on correlations, including experiments Reko-3 of FzJ, despite if forced flow. The coupling with MELCOR 1.8.5 allows to follow the evolutionary dynamics of the atmosphere and has been validated by experiments in the series Zx at Battelle Model Containment. Finally, it has demonstrated the feasibility of coupling model with type CFD codes by implementing CFX-4.4 the intercomparison exercise PARIS codes. Despite the absence of specific experiments, another interesting finding of this study is to analyze the behavior of the device in the presence of CO, showing an adequate prediction of the processes of adsorption and transport of species in particular circumstances.