Inter-ELM evolution of the edge current density profile on the ASDEX Upgrade tokamak

The sudden decrease of plasma stored energy and subsequent power deposition on the first wall of a tokamak device due to edge localised modes (ELMs) is potentially detrimental to the success of a future fusion reactor. Understanding and control of ELMs is critical for the longevity of these devices and also to maximise their performance. The commonly accepted picture of ELMs posits a critical pressure gradient and current density in the plasma edge, above which coupled magnetohydrodynamic (MHD) peeling-ballooning modes are driven unstable. Much analysis has been presented in recent years on the spatial and temporal evolution of the edge pressure gradient. However, the edge current density has typically been overlooked due to the difficulties in measuring this quantity. In this thesis, a novel method of current density recovery is presented, using the equilibrium solver CLISTE to reconstruct a high resolution equilibrium utilising both external magnetic and internal edge kinetic data measured on the ASDEX Upgrade (AUG) tokamak. The evolution of the edge current density relative to an ELM crash is presented, showing that a resistive delay in the buildup of the current density is unlikely. An uncertainty analysis shows that the edge current density can be determined with an accuracy consistent with that of the kinetic data used. A comparison with neoclassical theory demonstrates excellent agreement between the current density determined by CLISTE and the calculated profiles. Three ELM mitigation regimes are investigated: Type-II ELMs, ELMs suppressed by external magnetic perturbations (MPs), and Nitrogen seeded ELMs. In the first two cases, the current density is found to decrease as mitigation onsets, indicating a more ballooning-like plasma behaviour. In the latter case, the flux surface averaged current density can decrease while the local current density increases, thus providing a mechanism to suppress both the peeling and ballooning modes.