Une simple classification morphologique des galaxies de l'Univers local montre deux grandes familles: (1) les galaxies disques, avec des bras spiraux et dans deux-tiers des cas une barre stellaire; et (2) les galaxies elliptiques et lenticulaires, dites galaxies de type précoce ou early-type galaxies (ETGs), qui sont dominées par une composante stellaire sphéroidale. Les galaxies les plus massives de l'Univers local sont les ETGs. Ces galaxies présentent aussi une large variété de dynamique stellaire: elles peuvent avoir un champ de vitesse régulier et aligné avec la photométrie ou perpendiculaire à la photométrie; ne présenter aucune rotation globale; ou alors être composées de deux disques en contre-rotation l'un par rapport à l'autre (Kinematically Distinct Core ou KDC). Ces signatures dans la dynamique stellaire des ETGs et leur importante masse sont des signes d'interactions passées, en particulier des signes de fusions de galaxies. Le but principal de ma thèse est d'analyser un large échantillon de simulations numériques à haute résolution de fusions binaires de galaxies. Ces fusions sont dites "idéalisées" car elles ne prennent pas en compte le contexte cosmologique de formation des galaxies : deux galaxies en isolation sont lancées sur une orbite permettant la fusion de ces galaxies, le résultat final attendu de la fusion étant une ETG. L'analyse statistique de ce large échantillon de simulations nous permet de relier les conditions initiales de la fusion à la galaxie finale. J'ai démontré que le rapport de masse entre les spirales initiales et que l'orientation de leurs moments angulaires sont des points essentiels à la formation des ETGs avec peu ou beaucoup de rotation et des KDCs. La morphologie de la spirale (rapport Bulbe/Disque) est aussi un point important pour la formation des KDC mais son impact n'est pas clair et de nouvelles simulations sont nécessaires pour conclure. Durant ma thèse, j'ai aussi étudié l'importance de la résolution dans les simulations numériques de fusion de galaxies. J'ai montré que le nombre de particules et la taille des cellules utilisées ont une importance prépondérante dans les résultats finaux. Une trop faible résolution (i.e. peu de particules et une grille grossière) ne permet pas de suivre l'évolution rapide du potentiel gravitationnel lors de la fusion. Dans ce cas, certaines particules n'évacuent pas leur moment angulaire vers l'extérieur de la galaxie: la galaxie résultante de la fusion garde ainsi une plus forte rotation. A haute résolution, la dispersion de ces orbites est résolue, la galaxie résultante possède donc une faible rotation et peut former, sous certaines conditions initiales, un KDC.

<p>In order to reduce the discharge of the greenhouse gas CO2, the use of biomass is nowadays promoted as fuel in boilers. Compared to boilers fired with coal and oil the biomass-fired boilers have more complications related to both fouling and corrosion on the heat transfer surfaces. After the combustion, unburned inorganic matter in state of vapour, melts and solid particles are transported in the flue gas and may form deposits on heat transfer surfaces.</p><p>Deposits on the heat transfer surfaces may result in both increasing corrosion and decreasing boiler efficiency as the heat transfer rate to the superheaters and reheaters decrease by deposits.</p><p>In order to understand the process of deposit build-up, the whole combustion and transport process had to be analysed including aspects such as, boiler design, fuel properties and combustion environment, followed by particle transport phenomena and the probability for particles to get stuck on the heat transfer tubes.</p><p>In this thesis numerical simulation of particle trajectories has been conducted as well as measurements of deposits on a special designed deposit probe followed by investigation of on-site measurements of deposit depth on the super-heater tubes in a circulating fluidised bed in Västerås, Sweden.</p><p>Numerical simulations of particle trajectories in the vicinity of two super-heater tubes were conducted in an Eulerian-Lagrangian mode considering the flue gas and ash particles phase. Particle impingements on the tubes were investigated for different particle sizes. The results from the particle trajectory simulations show that particle larger than 10 µm will mainly impinge on the windward side of the first tube but, however also on the sides of the second tube in the flue gas flow direction. In theory as well as from observations and measurements two tubes can merge together by the deposit build-up. Smaller particles are usually more dispersed due to turbulence and thermophorectic forces, resulting in a more even impingement distribution on the whole surface of the tubes.</p><p>Probe measurements reveal that the deposit layer growth rate have a significant temperature and time dependence. After the initial deposit build-up a sintering process occurs and sintering is also proven to be dependent on temperature and exposure time.</p><p>Soot-blowing is the most common method to reduce the effect of deposits on the heat transfer tubes. In the present thesis the soot boiling efficiency is therefore also investigated. The soot-blowing show a strong positive effect on the heat transfer rate in a short time (hours) perspective after a soot-blowing cycle is completed. This positive effect is much weaker when considering a time period of three years. This is an effect of fact that soot-blowing mostly remove the loose part of the deposit material leaving the hard sintered part unaffected.</p><p>The subject of deposit build up on superheater tubes in large scale boilers involves multi-discipline knowledge and historically, the related research is mostly conducted as measurements and experiments on operating plants. Possibly in the future, theoretical simulations will have a bigger part of research on deposit build-up where the calculations are to be calibrated through measurements on real sites plants.</p>

<p>This paper examines the pricing of barrier options using Monte Carlo Simulations. MATLAB based software is developed to estimate the price of the option using Monte Carlo simulation. We consider a generalized barrier option of knock out type, but we let the domain take the shape of a rectangular box. We investigate the price of this kind of barrier options. We investigate how the box is placed and what effect it will have on the price of the option. We compare the number of trajectories that are needed in order to achieve the same accuracy between this box barrier option and an ordinary option.</p>

<p>In today’s industrial world, there are high demands on the environmental aspects.</p><p>Siemens Industrial Turbomachinery AB (SIT AB) is a company that is keen about the environment, and therefore spends a lot of effort in developing combustion processes in order to reduce NOx (nitrogen oxides) emissions on their engine products. They are also researching in optional fuels, which are more environment-friendly.</p><p>In order to provide lower emissions the SIT designed a water rig to study the flow dynamics in a DLE (Dry Low Emission) burner.</p><p>An analyze program (GUI horizontal) was developed with new functions and the existing functions were improved. The program’s function was to evaluate different experimental tests of the flow dynamics in the 3rd generation DLE burners, of the SGT-800 gas turbine engine.</p><p>The aim was to ensure repeatability to enhance reliability, of the experimental test results for further comparison, for upcoming projects concerning future DLE burners.</p><p>When repeatability was achieved, implementations of different geometrical modifications were performed in the 3rd generation DLE burner.</p><p>The reason of the geometrical alterations was to look over if better fuel air mixture could be obtained and accordingly (thus) to reduce hotspots in the burner and in that case reduce NOx emissions.</p>

<p>Decision making with the help of graphs has been applied for a long time. Previous to the introduction of computerized solutions, graphs and complementary pictures were hand drawn. With the introduction of computers came the colored and dynamic “animations” called simulations which are used today. The focal company of this study, Ångpanneföreningen AB, has during an extensive period of time constructed simulations of complex systems which they have used in order to satisfy specific needs of their customers in different areas. Depending on the complexity of the product and the needs it is ought to fulfill, the exchange of it will have a direct impact on the interactions carried out between the buying and selling parties in various ways. A question that this thesis looked to answer was how the utilization of simulations affects Ångpanneföreningen AB’s relationship with its customers. The purpose of this study is to describe how the use of a simulation program may affect the relation between two parties in a business relationship. A case study approach based on focused and semi-structural interviews has been used for this study.The effects of the use of simulations on a business relationship, in terms of benefits perceived, depend on the purpose set for the software. This meaning that it depends on whether the results and benefits are of such kind that they are possible to measure. However, if measurable, then the awareness of perceived benefits contributes to a stronger relationship since trust is likely to follow from met expectations. The simulation software offered by ÅF has the characteristics of both a product and a service with a highly technological nature. When dealing with simulations holding such attributes, a high level of exchanges is likely to be involved. Furthermore, interdependencies between two parties in a business relationship tend to develop as both the exchanges due to, and the complexities of, the simulation are of a high level, meaning that exchanges are needed in both directions.</p>

<p>Fixed-collision correlated sampling for Monte Carlo (MC) simulations is a method which can be used in order to shorten the simulation time for brachytherapy treatment planning in a 3D patient geometry. The increased efficiency compared to conventional MC simulation is measured by efficiency gain. However, a previous study showed that, in some cases, PDFs (probability density functions) of estimates of the efficiency gain, simulated using resampling and other MC methods, were multimodal with values below 1. This means that the method was less effective than conventional sampling for these cases. The aims of this thesis were to trace the causes of the multimodal distributions and to propose techniques to mitigate the problem caused by photons with high statistical weights.Two simulation environments were used for the study case, a homogeneous and a heterogeneous environment. The homogenous environment consisted of a water sphere with the radius 100mm. For the heterogeneous environment a cylindrical block of tungsten alloy (diameter 15 mm, height 2.5 mm) was placed in the water sphere. The sphere was divided into an array of cubic voxels of size 2.5 mm x 2.5 mm x 2.5 mm for dose calculations. A photon source was positioned in the middle of the water sphere and emitted photons with the energy 400 keV.It was found that the low values and multimodal PDFs for the efficiency gain estimates originated from photons depositing high values of energy in some voxels in the heterogeneous environment. The high energy deposits were due to extremely high statistical weights of photons interacting repeatedly in the highly attenuating tungsten cylinder. When photon histories contributing to the rare events of high energy deposits (outliers) were removed, the PDFs became uni-modal and efficiency gain increased. However, removing outliers will cause results to be biased calling for other techniques to handle the problem with high statistical weights.One way to resolve the problem in the current implementation of the fixed-collision correlated sampling scheme in PTRAN (the MC code used) could be to split photons with high statistical weights into several photons with the same sum weight as the initial photon. The splitting of photons will result in more time consuming simulations in areas with high attenuation coefficients, which may not be the areas of interest. This could be resolved by using Russian roulette, eliminating some of the photons with high statistical weight in such areas.Fixed-collision correlated sampling for Monte Carlo (MC) simulations is a method which can be used in order to shorten the simulation time for brachytherapy treatment planning in a 3D patient geometry. The increased efficiency compared to conventional MC simulation is measured by efficiency gain. However, a previous study showed that, in some cases, PDFs (probability density functions) of estimates of the efficiency gain, simulated using resampling and other MC methods, were multimodal with values below 1. This means that the method was less effective than conventional sampling for these cases. The aims of this thesis were to trace the causes of the multimodal distributions and to propose techniques to mitigate the problem caused by photons with high statistical weights.Two simulation environments were used for the study case, a homogeneous and a heterogeneous environment. The homogenous environment consisted of a water sphere with the radius 100mm. For the heterogeneous environment a cylindrical block of tungsten alloy (diameter 15 mm, height 2.5 mm) was placed in the water sphere. The sphere was divided into an array of cubic voxels of size 2.5 mm x 2.5 mm x 2.5 mm for dose calculations. A photon source was positioned in the middle of the water sphere and emitted photons with the energy 400 keV.It was found that the low values and multimodal PDFs for the efficiency gain estimates originated from photons depositing high values of energy in some voxels in the heterogeneous environment. The high energy deposits were due to extremely high statistical weights of photons interacting repeatedly in the highly attenuating tungsten cylinder. When photon histories contributing to the rare events of high energy deposits (outliers) were removed, the PDFs became uni-modal and efficiency gain increased. However, removing outliers will cause results to be biased calling for other techniques to handle the problem with high statistical weights.One way to resolve the problem in the current implementation of the fixed-collision correlated sampling scheme in PTRAN (the MC code used) could be to split photons with high statistical weights into several photons with the same sum weight as the initial photon. The splitting of photons will result in more time consuming simulations in areas with high attenuation coefficients, which may not be the areas of interest. This could be resolved by using Russian roulette, eliminating some of the photons with high statistical weight in such areas.</p>

<p>The use of process capability indices in the industry is traditionally based on the assumptions that a sample from a process are observations on independently, identically and normally distributed random variables<i>, IIN</i>. However, all three assumptions are open to discussion and in this thesis, the estimation of the indices is studied when the assumption of independence is not fulfilled.</p><p>In five reports, the indices <i>C</i><i>pk</i> and <i>C</i><i>pm </i>are studied, and instead of random samples, samples are regarded as observations on a time series.</p><p>In the first four reports, each index is studied for either an <i>AR(1)</i> or an <i>MA(1)</i> process and the fifth report, both indices are studied for a general <i>ARMA(p,q</i>) process.</p><p>In all reports, alternatives to <i>C</i><i>pk</i><i> </i>and <i>C</i><i>pm</i><i> </i>are suggested as well as point and interval estimators for the suggested indices. The accuracy of interval estimators are evaluated through large Monte Carlo simulations and the difference between empirical coverage rates and nominal confidence limits are calculated.</p><p>It was found in all reports that a dependency among observations has a great impact on the coverage rates. The coverage rate difference depends on both the size of the autocorrelation and the type of time series model and for the original <i>C</i><i>pk</i> and <i>C</i><i>pm</i> the difference can be large. With the suggested alternative indices, however, the differences are always reduced and unless the autocorrelations are close to ±1, the sizes of differences are of little consequence.</p>

Composite materials reinforced with particles or fibers are widely used in industrial applications due to their good mechanical, thermal, and electrical properties. Consequently, for the scientific community as well as the industry, an important challenge is to understand the relationship between the microstruture and the macroscopic response in order to design composite materials with optimised properties. In this thesis, we study a class of inclusion-reinforced multi-phase composites. Our main objective is to develop a micromechanical model and the corresponding numerical algorithms which enable the simulation of the rate-independent mechanical response. The proposed model is based on an incremental Hill-type formulation and uses the two-step Mori-Tanaka/Voigt mean-field homogenisation schemes. The crucial issues of the choice of reference comparison materials and Eshelby's tensor computation are examined In parallel, an experimental study consisting in four-point bending tests performed on plain concrete and steel fiber-reinforced concrete (SFRC) specimens, is carried out with the aim of achieving an appropriate modelling of SFRC, and collecting data for the validation of our model predictions. The accuracy and the efficiency of the proposed approach are evaluated through numerical simulations. Several discriminating tests of concrete, metal, and polymer matrix composites are carried out. A two-scale approach is developed in order to simulate, within reasonable CPU time and memory usage, the response of realistic structures under complex loadings. In many cases our estimates are validated against finite element computations and experimental results.

Recent advances in the fundamental theories in cryobiology using thermodynamic principles have created new opportunities for innovative methodologies in cryobiology. This thesis tested the hypothesis that calculated indicators of the two-factor hypothesis of cryoinjury, depending on cellular osmotic properties, will describe outcomes of cryobiological experiments. In addition, this thesis demonstrated that knowledge gained from improved descriptions of cellular osmotic parameters allows better understanding of cryoinjury and cryoprotection. The main objective of this thesis was to develop approaches using simulations that can be applied to development of cryopreservation procedures for cell types of interest for therapies. In order for this approach to be successful, a method to more accurately describe the osmotic solution properties of the cell (i.e. osmolality as a function of molality for the cytoplasm) was developed. Also, in-depth examination into the correlation between predictions of the two types of cryoinjury and measured post-thaw biological outcomes was required. The work presented in this thesis has shown that simulations, based on cell-specific osmotic characteristics, and coupled with interrupted cooling procedures can be used to determine conditions that minimize the two identified damaging factors in cryopreservation. Based on results from this research, both intracellular supercooling and osmolality, as indicators of intracellular ice formation and solution effects injury, respectively, should be calculated when attempting to compare simulations with biological experimentation. This thesis has also shown a novel method of obtaining the solution properties (i.e. osmolality as a function of molality) of the cytoplasm of living cells using equilibrium cell volume measurements. Using these newly calculated parameters, this research also demonstrated the magnitude of error introduced by making dilute solution assumptions of the solution properties in cellular responses to low temperatures, including simulations of interrupted freezing procedures. Overall, the research work presented in this thesis has extended the approach to cryopreservation to include the properties of the cell and the physical conditions of the freezing environment, which was only possible through the linkage between biological experimentation and simulations.

While rheological and microstructural complexities have posed tremendous challenges to researchers in developing first principles models and simulation techniques that can accurately and robustly predict the dynamical behaviour of polymeric flows, the past two decades have offered several significant advances towards accomplishing this goal. These accomplishments include: (1). Stable and accurate formulation of continuum-level viscoelastic constitutive models and their efficient implementation using operator splitting methods to explore steady and transient flows in complex geometries, (2). Prediction of rheology of polymer solutions and melts based on micromechanical models as well as highly parallel self-consistent multiscale simulations of non-homogeneous flows. The main objective of this study is to leverage and build upon the aforementioned advances to develop a quantitative understanding of the flow-micro-structure coupling mechanisms in viscoelastic polymeric fluids and in turn predict, consistent with experiments, their essential macroscopic flow properties e.g. frictional drag, interface shape, etc. To this end, we have performed extensive continuum and multiscale flow simulations in several industrially relevant bulk and free surface flows. The primary motivation for the selection of the specific flow problems is based on their ability to represent different deformation types, and the ability to experimentally verify the simulation results as well as their scientific and industrial significance.