Le Chauffage Cyclotron Ionique (ICRH) par des ondes dans la gamme 30-80MHz est couramment utilisé dans les plasmas de fusion magnétique. Excitées par par des réseaux phasés de rubans de courant à la périphérie du plasma, ces ondes existent sous deux polarisations. L’onde rapide traverse le bord ténu du plasma par effet tunnel puis se propage à son centre où elle est absorbée. L’onde lente, émise de façon parasite, existe seulement à proximité des antennes. Quelle puissance peut être couplée au centre avec 1A de courant sur les rubans? Comment les champs radiofréquence (RF) proches et lointains émis interagissent-ils avec le plasma de bord par rectification de gaine RF à l’interface plasma-paroi? Pour répondre simultanément à ces deux questions, en géométrie réaliste sur l’échelle spatiale des antennes ICRH, cette thèse a amélioré et testé le code numérique SSWICH (Self-consitent Sheaths and Waves for ICH). SSWICH couple de manière auto-cohérente la propagation des ondes RF et la polarisation continue (DC) du plasma via des conditions aux limites non-linéaires de type gaine (SBC) appliquées à l’interface plasma / paroi. La nouvelle version SSWICH-FW est pleine onde et a été développée en deux dimensions (toroïdale/radiale). De nouvelles SBCs couplant les deux polarisations d’ondes ont été obtenues et mises en œuvre le long de parois courbes inclinées par rapport au champ magnétique de confinement. Avec ce nouvel outil en l'absence de SBCs, nous avons étudié l'impact d'une densité décroissant continûment à l'intérieur de la boîte d'antenne en traversant la résonance hybride basse (LH). Dans les limites mémoire de notre poste de travail, les champs RF au-dessous de la résonance LH ont changé avec la taille de maille. Par contre spectre de puissance couplée n’a que très peu évolué, et n’était que faiblement influencé par la densité à l'intérieur de l'antenne. En présence de SBCs, les simulations SSWICH-FW ont identifié le rôle de l'onde rapide sur l’excitation de gaines RF et reproduit certaines observations expérimentales clés. SSWICH-FW a finalement été adapté pour réaliser les premières simulations 2D électromagnétiques et de gaine-RF de la machine plasma cylindrique magnétisée ALINE / Ion Cyclotron Resonant Heating (ICRH) by waves in 30-80MHz range is currently used in magnetic fusion plasmas. Excited by phased arrays of current straps at the plasma periphery, these waves exist under two polarizations. The Fast Wave tunnels through the tenuous plasma edge and propagates to its center where it is absorbed. The parasitically emitted Slow Wave only exists close to the launchers. How much power can be coupled to the center with 1A current on the straps? How do the emitted radiofrequency (RF) near and far fields interact parasitically with the edge plasma via RF sheath rectification at plasma-wall interfaces? To address these two issues simultaneously, in realistic geometry over the size of ICRH antennas, this thesis upgraded and tested the Self-consistent Sheaths and Waves for ICH (SSWICH) code. SSWICH couples self-consistently RF wave propagation and Direct Current (DC) plasma biasing via non-linear RF and DC sheath boundary conditions (SBCs) at plasma/wall interfaces. Its upgrade is full wave and was implemented in two dimensions (toroidal/radial). New SBCs coupling the two polarizations were derived and implemented along shaped walls tilted with respect to the confinement magnetic field. Using this new tool in the absence of SBCs, we studied the impact of a density decaying continuously inside the antenna box and across the Lower Hybrid (LH) resonance. Up to the memory limits of our workstation, the RF fields below the LH resonance changed with the grid size. However the coupled power spectrum hardly evolved and was only weakly affected by the density inside the box. In presence of SBCs, SSWICH-FW simulations have identified the role of the fast wave on RF sheath excitation and reproduced some key experimental observations. SSWICH-FW was finally adapted to conduct the first electromagnetic and RF-sheath 2D simulations of the cylindrical magnetized plasma device ALINE

QC 20100622

<p>NR 20140805</p>

No description available.

This thesis concerns heating and current drive in tokamak plasmas using the fast magnetosonic wave in the ion cyclotron range of frequencies. Fast wave heating is a versatile heating method for thermonuclear fusion plasmas and can provide both ion and electron heating and non-inductive current drive. Predicting and interpreting realistic heating scenarios is however difficult due to the coupled evolution of the cyclotron resonant ion velocity distributions and the wave field. The SELFO code, which solves the coupled wave equation and Fokker-Planck equation for cyclotron resonant ion species in a self-consistent manner, has been upgraded to allow the study of more advanced fast wave heating and current drive scenarios in present day experiments and in preparation for the ITER tokamak. Theoretical and experimental studies related to fast wave heating and current drive with emphasis on fast ion effects are presented. Analysis of minority ion cyclotron current drive in ITER indicates that the use of a hydrogen minority rather than the proposed helium-3 minority results in substantially more efficient current drive. The parasitic losses of power to fusion born alpha particles and beam injected ions are concluded to be acceptably low. Experiments performed at the JET tokamak on polychromatic ion cyclotron resonance heating and on fast wave electron current drive are presented and analysed. Polychromatic heating is demonstrated to increase the bulk plasma ion to electron heating ratio, in line with theoretical expectations, but the fast wave electron current drive is found to be severely degraded by parasitic power losses outside of the plasma. A theoretical analysis of parasitic power losses at radio frequency antennas indicates that the losses can be significantly increased in scenarios with low wave damping and with narrow antenna spectra, such as in electron current drive scenarios. / QC 20100506

In a fusion plasma, the ions have to be heated to reach fusion relevant temperatures. One possibility is to launch an electromagnetic wave in the radio frequency band into the plasma. This wave can resonate with the ions at their cyclotron frequency and hence, the method is called ion cyclotron resonance heating. If the Larmor radius is of similar length scales as the wavelength, finite Larmor radius (FLR) effects are important. This introduces additional possible wave modes. To accurately predict the heating, simulations of these additional modes can be important. In order to describe the FLR effects in simulations, this work applies a Taylor expansion to second order in the perpendicular wavenumber of the dielectric tensor. In real space, the Taylor expansion corresponds to a series of spatial derivatives. These derivatives are implemented as weak-form contributions for a onedimensional finite element method (FEM)simulation. The approach is realized in a fork of the FEMIC code, coupling plasma physics in MATLAB with the FEM solver in COMSOL. The discretization of the FEM solver is adapted using a Helmholtz filter to provide the required degree of smoothness. Our approach proves successful to simulate FLR effects within the limitations caused by the modeling choices. We compare results with an all-order FLR method, showing good qualitative agreement. This work serves as a proof of concept to describe challenges on the way towards incorporation of second order FLR effects in twodimensional simulations in FEMIC. / I ett fusionsplasma måste jonerna värmas för att plasmat ska nå temperaturer relevanta för fusion. En möjlighet är att sända in en elektromagnetisk våg i radiofrekvensbandet i plasmat. Den vågen kan sedan resonera med jonerna vid deras cyklotronfrekvens, och därför kallas metoden för joncyklotronresonansuppvärmning. Om våglängden är jämförbar med Larmorradien blir ändliga Larmorradie-effekter viktiga. Detta möjliggör ytterligare typer av vågor. Det kan vara viktigt att simulera dessa typer av vågor för att förutsäga uppvärmningen på ett träffsäkert vis. I denna masteruppsats Taylorutvecklar vi den dielektriska tensorn till andra ordningen i det vinkelräta vågtalet för att beskriva hur FLR-effekter påverkar simuleringarna. I det reella rummet motsvarar Taylorutvecklingen en serie av rumsliga derivator. Dessa derivator implementeras sedan som bidrag på svag form i en endimensionell modell som löses med den finita elementmetoden (FEM). Metoden implementeras i FEMIC-koden, som kopplar plasmafysik i MATLAB med FEM-lösaren i COMSOL. Diskretiseringen av FEM anpassas med ett Helmholtzfilter för att få en tillräckligt slät funktion. Ansatsen visar sig kunna framgångrikt simulera FLR-effekter, med vissa förväntade begränsningar. Lösningen jämförs sedan med lösningen från en metod som tar hänsyn till FLR-effekter, men som inte är baserad på en serieuteckling. Vi finner god kvalitativ överensstämmelse. Detta arbete fungerar som en prototyp och ämnar att beskriva de utmaningar som kan uppstå vid implementation av FLR-effekter den tvådimensionella axisymmetriska versionen av FEMIC.