The Rayleigh-Taylor instability (RTI) is one of the primary hydrodynamic instabilities that acts as a disputer to achieving high yield inertial confinement fusion (ICF). The potential for RTI to grow on the interior surface of ICF capsules, caused by deceleration during the implosion, further emphasises the need to better understand the seed mechanisms for RTI and possible mitigation methods for damping the instability growth. Reducing the growth of RTI during deceleration could preserve the spherical symmetry of ICF implosions and reduce the amount of mix between the solid capsule liner and fuel hot-spot. Additionally, it has been shown that magnetic fields do damp RTI growth, and the presence of a magnetic field lowers the threshold for achieving fusion and increases the yield. Understanding the seed mechanisms of the RTI, especially on the interior surface of ICF capsules, further allows for better understanding of the morphology of the RTI growth dur- ing deceleration. Classically RTI has been studied using single or multi-mode sinusoidal perturbations, which result in bubble and spike morphology. However in addition to si- nusoidal perturbations, single-feature perturbation, such as voids or divots, can seed RTI. This form of RTI is considered the thin-layer RTI, where the perturbation's wavelength is longer than the dense layer's thickness. This specific RTI evolution results in a morphology consisting of a single central spike and arms that extend horizontally away from the spike and eventually fall back towards the interface. Thin-layer RTI is important to explore dur- ing deceleration due to the presence of the fill-tubes in ICF capsules causing holes in the shell. Creating experimental platforms for current laser configurations on Omega and the Na- tional Ignition Facility (NIF) is necessary to study deceleration-stage RTI experimentally and validate computational modeling. A comprehensive exploration of potential experimen- tal designs on Omega, Omega-EP, and NIF are explored to identify a platform with which deceleration-stage RTI can be studied with and without the presence of an externally applied magnetic field. Additionally, the design of a novel experimental platform for Omega-EP to study thin-layer RTI during deceleration with and without an externally applied magnetic field is presented, along with data collected during the first experiments performed utilizing the platform. Lastly, a first of it's kind RTI platform for NIF is fielded and the results are presented, including an exploration of the possible impacts high-intensity-laser generated hot-electrons can have on experimental targets. The results of these experimental platforms are used to benchmark computational models, and demonstrate the potential for magnetized RTI to be studied comprehensively in future experiments. / Doctor of Philosophy / The potential of controlled sustained nuclear fusions as a viable energy source has rapidly become a reality in recent years. Monumental progress has been made in the pursuit of con- trolled fusion, including the repeated achievement of fusion ignition at the National Ignition Facility (NIF), meaning there was successful production of more energy from the fusion reac- tion than laser energy used to trigger the reaction. However, in order for fusion to become a truly viable energy source improvements in capsule design and the mitigation of disruptions, like hydrodynamic instabilities, must be explored to produce higher energy yields. The Rayleigh-Taylor instability (RTI) is one of the most detrimental hydrodynamic insta- bilities in inertial confinement fusion (ICF). RTI occurs when a lighter fluid, like the fuel used in fusion reactions, supports a heavier fluid, the ICF capsule itself, under the influence of gravity. An ICF capsule is imploded, induced by the driving mechanism, such as a laser, but once the driver stops the capsule will begin to decelerate. During this deceleration stage, the interior surface of the ICF capsule in susceptible to RTI growth causing the cold capsule material to mix with the hot fusion fuel. This mixing reduces the fuel's ability to reach the necessary temperatures and densities need to achieve ignition and produce high energy yields. As a result, it is crucial to better understand the defects that cause RTI to grow and explore methods that could damp the RTI growth and preserve the integrity of the implosion and fusion fuel. The work presented here focuses on exploring both the seed mechanisms for RTI and miti- gation strategies. Specifically, using an externally applied magnetic field has been shown to damp RTI growth and in know to lower the threshold of the conditions needed to achieve ignition. A study of possible experimental setups at both the Omega laser and NIF is ex- plored in order to identify a design with which the damping effects of an externally applied magnetic field on deceleration-stage RTI can be studied experimentally. From this design study platforms for the Omega-EP and NIF were conceptualized and ultimately fielded. The results from these novel experiments are presented, along with an exploration of pos- sible effects on RTI unexpected preheating of the experimental targets. Additionally, an exploration of the seed mechanisms of RTI is presented with a look at the classic sinusoidal perturbation as well as using a divot to seeded thin-layer RTI, which evolves with a spike and arm morphology rather than the classical bubble and spike. The experimental results from Omega-EP using a divot as the perturbation are presented. Novel results of varying RTI platforms and their potential for further development provide crucial insight into the possible presence of deceleration-stage RTI in ICF capsules and can be iterated on in the future to further explore RTI evolution and damping methods.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Pedestrian detection is a canonical instance of object detection that remains a popular topic of research and a key problem in computer vision due to its diverse applications. These applications have the potential to positively improve the quality of life. In recent years, the number of approaches to detecting pedestrians in monocular and binocular images has grown steadily. However, the use of multispectral imaging is still uncommon. This thesis work presents a novel approach to data and feature fusion of a multispectral imaging system for pedestrian detection. It also includes the design and building of a test rig which allows for quick data collection of real-world driving. An application of the mathematical theory of trifocal tensor is used to post process this data. This allows for pixel level data fusion across a multispectral set of data. Performance results based on commonly used SVM classification architectures are evaluated against the collected data set. Lastly, a novel cascaded SVM architecture used in both classification and detection is discussed. Performance improvements through the use of feature fusion is demonstrated.

Neutron emission spectra, measured with the time of flight spectrometer TOFOR, at the joint European torus (JET) are presented in this thesis. TOFOR has been in use since 2005, routinely measuring the neutron emission from JET plasmas. The work in the thesis mainly concerns the modeling of the signatures in the neutron spectrum that reveal different parts of the fuel ion distribution, such as the thermal bulk plasma as well as energetic ions from neutral beam and ion cyclotron heating. Parametric models of the signatures, using plasma parameters as input, are employed to generate trial neutron spectra. The parameters, such as the fuel ion temperature or the fast ion distribution function, are deduced by iteratively fitting the trial spectra to the measured data. Measurements with TOFOR have been made and the models were applied. The studies are mainly on neutrons from d(d, n)3 He reactions(DD), although the emission from reactions with the plasma impurity 9 Be and triton burn up is covered as well. This has allowed for detailed studies of e.g. the physics ICRF heating as well as the interactions between energetic ions and plasma instabilities, such as toroidal Alfvé Eigenmodes.

Diagnostic measurements play a vital role in experiments. Without them we would be in the dark with no way of knowing what was happening; of understanding the processes and behaviour occurring; or even of judging the success or failure of our experiments. The development of fusion plasma devices is no different. In this thesis we concentrate on visible spectroscopy based diagnostics: examining the techniques for measurement and analysis; the breadth of plasma parameters that can be extracted from the spectroscopic data; and how the application of these diagnostic techniques gives us a broader picture of the plasma and the events taking place within. Techniques are developed and applied to plasmas in three fusion experiments, EXTRAP T2R, ASDEX Upgrade and JET. The diagnostic techniques exploit different features of the measurements of the emitted photons to obtain various useful plasma parameters. Determination of the ion temperature and rotation velocity of oxygen impurity ions in the EXTRAP T2R plasma is achieved through measurement and analysis of, respectively, the Doppler broadening and the Doppler wavelength shift of visible wavelength atomic spectral lines. The evolution of the temperature and rotation is studied as a function of the discharge parameters, in particular looking at the effect of applying active feedback control schemes to the resistive wall modes and/or pulsed poloidal current drive. Measurements of multiple ionisation stages are used to estimate radial profiles of the toroidal rotation and the ion temperature and correlations between the ion rotations and the rotation velocities of tearing modes are also established. Radial profiles of the emissivity and density (or concentration) of the oxygen ions are obtained by means of measurements of the spectral line intensities on a small array of linesof- sight through the plasma. Changes to the profiles for different plasma schemes and the implications for particle transport are investigated. The derived emissivity profiles are used in the analysis for some of the other spectroscopic diagnostics. Spectral line intensity measurements (in this case of neutral ions) are also the basis for calculations of both the electron temperature and the particle fluxes at the plasma edge. The latter is an indicator of the degree and type of interaction between the plasma and the surrounding surfaces. Particle fluxes of the operating gas hydrogen and of chromium and molybdenum impurities are investigated in EXTRAP T2R for different operating scenarios, in particular changes in the metallic influx with the application of active feedback mode control are examined along with the correspondence between spectroscopic and collector probe results. In the ASDEX Upgrade divertor estimates of the particle flux of the deuterium operating gas are also made through analysis of spectral intensities. Molecular D2 band structure is explored in addition to the Balmer Dα spectral line intensity to acquire both atomic and molecular particle fluxes, investigate the contribution of the dissociating D2 to the Dα line and study the effect of changes in the divertor. Analysis of the D2 molecular band structure (the relative intensities of the rotational lines and vibrational bands) also enables calculation of the upper state rotational and ground state vibrational temperatures. The locations of emitting atomic ions in JET are estimated from Zeeman splitting analysis of the structure of their spectral lines. The measurement and analysis of visible wavelength light is demonstrated to be a sensitive diagnostic tool in the quest for increased knowledge about fusion plasmas and their operating scenarios. / QC 20100810

Understanding of plasma confinement modes is an essential component in development of a fusion reactor. Plasma confinement directly relates to performance of a fusion reactor in terms of energy replacement time requirements on other design parameters. Although a variety of levels of confinement have been achieved under different operating conditions in tokamaks, tokamak confinement is generally identified as being either Low (L-mode--poor confinement) or High (H-mode--good confinement) In operation of a tokamak experiment, the plasma confinement condition generally changes from L-mode to H-mode over a few hundred milliseconds, sometimes quite sharply. Such a difference in transition period seems to be largely due to operating conditions of the plasma. Comparison of experimental data exhibits various distinctions between confinement modes. One noteworthy distinction between confinement modes is development of steep density and temperature gradients of electrons and ions in the plasma edge region of High confinement, H-modes, relative to Low-confinement, L-modes. The fundamental reason for the change for L-mode to H-mode is not understood. Previous studies have suggested i) the development of reduced diffusive transport coefficients that require a steepening of the gradients in a localized region in the edge plasma, the "transport barrier" in H-mode confinement ii) the radial force balance between pressure gradient forces and electromagnetic (radial electric field and VxB) forces require radial particle fluxes to satisfy a pinch-diffusion relation. A recent study suggests that the major difference between L-mode and H-mode are associated with the electromagnetic forces in the "pinch velocity" and the pressure gradient, not in the diffusion coefficients that multiplies the pressure gradient. The research will examine in detail the time evolution of the radial force balance and the particle and energy transport during the L-H transition. For the analysis, DIII-D shot #118897 is selected for transition between L- and H-mode confinements. Plasma conditions in L-mode, near the L-H transition and following the transition are selected for analysis of various parameter profiles. The initial analysis will be based on the four principal equations for plasma: particle balance, momentum balance, force balance and heat conduction. Based on these equations, specific equations have been derived: toroidal and radial momentum balances, diffusion coefficient, pinch velocity and heat conduction relation for calculation of parameters. The analysis of these equations, using the measured data, will establish how various terms in the radial force balance (radial electric field, VXB (electromagnetic) force, and pressure gradient) and the diffusive transport coefficients evolve over the confinement mode transition.

The presence of a large pinch velocity in the edge pedestal of high confinement (H-mode) tokamak plasmas implies that particle transport in the plasma edge must be treated by a pinch-diffusion theory, rather than a pure diffusion theory. Momentum balance also requires the inclusion of a pinch term in descriptions of edge particle transport. A numerical investigation of solving generalized pinch-diffusion theory using methods extended from the numerical solution methodology of pure diffusion theory has been carried out. The generalized diffusion equation has been numerically integrated using the central finite-difference approximation for the diffusion term and three finite difference approximations of the pinch term, and then solved using Gauss reduction. The pinch-diffusion relation for the radial particle flux was solved directly and used as a benchmark for the finite-difference algorithm solutions to the generalized diffusion equation. Both equations are solved using several mesh spacings, and it is found that a finer mesh spacing will be required in the edge pedestal, where the inward pinch velocity is large in H-mode plasmas, than is necessary for similar accuracy further inward where the pinch velocity diminishes. An expression for the numerical error of various finite-differencing algorithms is presented.

Diss. (sammanfattning) Stockholm : Karol. inst., 2006. / Härtill 4 uppsatser.

Thèse : Gestion : Lille : 1976. / Multigraphié.

The need to make maximum use of renewable resources to the detriment of fossil fuels to achieve environmental goals with an increasing energy demand is driving research into the development of technologies to obtain energy from sources that are not currently being exploited, one of them being fusion energy. The aim of this report is to provide a general overview of fusion and to provide a critical opinion on whether fusion will become a commercial energy source in the future, and if so when. The followed methodology has been a literature review complemented by an interview to B Henric M Bergsåker, teacher and researcher at the KTH on fusion plasma physics and information person for the Swedish fusion research.In the results section the fusion physics and different technological approaches have been presented. Among the studied different projects, the ITER Tokamak magnetic reactor has been selected as the most promising of these projects, as a product of international collaboration, and it has been analyzed in more detail. The obtained results have been that fusion can be an inexhaustible, environmentally friendly and safe energy source. The first-generation fusion commercial reactors are expected to be part of the energy mix before 2100.