The numerical study of filament dynamics in tokamak scrape-off layer plasmas

Resumen

Filamentary transport has been experimentally observed in a multitude of magnetically confined fusion devices, especially of the tokamak type. Filaments are three dimensional field-aligned structures that make a pronounced appearance on the outboard mid-plane side of the device where the destabilising forces for their underlying generation mechanism are the greatest. Their basic mechanism is the interchange instability, comparable to Rayleigh-Bénard instability in fluids, and they have been modelled by a variety of 2D and 3D numerical models. Filaments are carriers of large quantities of particles and heat and as such, their presence in the SOL has implications for the target surface design in future fusion reactors. The intermittent nature of filamentary transport make them dangerous to the longevity of target surfaces and the eventually degrading plasma confinement. So in order to better understand their nature, numerical studies today try to replicate experimentally observed behaviour. However, the success in this has been limited and understanding the real nature of filamentary transport is essential. To this end, in this work, filaments have been studied as isolated structures first in the ‘seeded filament’ approach, where an individual filamentary structure is reconstructed and studied in isolation. The seeded filament approach is useful for parametric studies to understand the impact of individual parameters on filament dynamics. These simulations are performed using a 3D fluid edge model based on the Braginskii closures - this model is the TOKAM3X code. In this work, the code is operated on an isothermal, electrostatic mode to explore filamentary dynamics in edge plasmas. Special tools developed to track isolated and turbulent filamentary structures are used to later identify key properties of the filaments simulated using the TOKAM3X code. The impact of the parallel resistivity in the background plasma is an important aspect to understanding filamentary motion in the edge region. The resistivity in the plasma is, in general, a measure of the plasma’s ollisionality which in turn depends on the temperature at first order. But in reality, the collisionality depends on a ost of factors like recycling at the walls, and magnetic topology like X-points, that can electrically disconnect filamentary structures from the target surfaces. In the studies performed here, it is seen that the parallel resistivity of the plasma can change the velocity-size dynamics of the filament. A scan of filament velocity versus size reveal that they show similar scalings as those predicted by theory, i.e. inertial regime characterised by small filament ize, and sheath-connected regime characterised by larger filament size. It is seen here, that the parallel resistivity an modulate the transition between the two regimes if its value is varied. The impact of localised strong magnetic hear on seeded filament motion is also studied separately. These simulations show that magnetic shear if strong nough, can redistribute the filament density by altering the potential field structure associated to the filament and provoking poloidal transport of filamentary density. Further it is shown that sufficient filament density at the magnetic shear layer can provoke the generation of new, secondary filaments beyond the shear layer. Seeded filamentary motion is further validated against experimental observations. It is seen that seeded filamentary simulations are a good tool to understand filamentary dynamics, but replicating such dynamics is not easy. Seeded filament validation studies depend heavily on experimental data, and how well the numerical model approaches the dominant physics in the experiment. In this study, it is seen that the radial motion of the filament is a property that most numerical models reproduce with a good degree of fidelity. Apart from the physics simulated in the models, the initial conditions and the boundary conditions used have an impact on simulation outputs. The impact of different limiter geometries, for example, in this studies shows that they can have a strong impact upstream from the target surfaces, and thus the output of simulations. Comparing seeded filaments to experimental filaments involves comparing a system that is not completely turbulent (in the seeded filament simulation) to a completely turbulent scenario (experiment). In this work, turbulent filaments are generated using the TOKAM3X model, to then track and characterise the turbulent structures observed therein using the tools developed in this thesis. The results are further used to seed a filament based on the extracted properties of the turbulent filaments from the simulations. When the seeded filament case is compared to the turbulent filament simulations, it is seen that the average radial velocity of the turbulent filaments is recovered by the seeded filament simulation. But the average maximum velocity reported for turbulent filaments is not recovered by the seeded filament cases - comparing simulation with simulation (using the TOKAM3X model). Another outcome of studying turbulent filament simulations using TOKAM3X, has been the role of the separatrix on simulations. Most seeded filament studies do not physically simulate the separatrix. And the turbulent simulations show that the case without the separatrix does reproduce separatrix-like artefacts in the statistical properties of the transport. However, analyses of the simulations here show that differences do arise in the vicinity of the separatrix, when compared to the case without separatrix. These differences can have an impact on seeded filament studies when the filament is seeded close to the inner boundary.