Energy load on tokamak plasma-facing components and its dependence on plasma regimes

Abstract

The concept of nuclear fusion is seen as a possible low CO2 source for power generation which could one day replace nuclear fission power and fossil fuels. A key challenge in realizing a possible commercial fusion reactor is the proper handling of heat loads on plasma-facing components. This thesis presents an analysis of the energy loads on divertor and non-divertor components in the ASDEX Upgrade Tokamak (Garching, Germany). The impact of enhanced filament transport, which can dominate the radial convective transport in the scrape-off layer (SOL) to the first wall, is investigated using dedicated H-mode plasmas with a magnetic equilibrium characterized by a high upper triangularity δu and high elongation κ. The database features a variation in the safety factor q by changing the toroidal magnetic field Btor and a large coverage in the density operational space by changing the feed-forward flattop fuelling level on a shot-to-shot basis. Low density H-modes associated with typeI edge localised modes (ELMs) and high density H-modes associated with an increased radial convective transport have been acquired. The latter scenario is also known as quasicontinuous exhaust (QCE) regime. As a prerequisite for trustworthy energy load scalings in ASDEX Upgrade, the global energy balance has been established using the revised cooling water calorimetry. In addition to the revised calorimetry, contributions from plasma radiation and from neutral beam injection losses need to be considered to achieve a representative global energy balance. Based on the novel understandings, the task in achieving a reasonable global energy balance has been repeated at the JET Tokamak (Culham, United Kingdom). The experiments have revealed that the normalized energy loads of divertor and nondivertor components correlate linearly with the turbulence control parameter αt defining the strength of interchange drive. The transition from a high-density QCE regime to a Type-I ELMy H-mode scenario exhibits a consistent linear trend. A generalization of the achieved scaling has been realized by extending the database with ASDEX Upgrade discharges with different plasma characteristics. Simultaneously, substantial changes of up- and downstream SOL profiles have been observed. The width of the heat flux in the SOL undergoes a significant broadening reducing its peak heat flux and increasing the area on which the parallel heat flux is deposited. Furthermore, a reduction of the parallel conductive transport is proven by means of thermocouple measurements, although not fully consistent with similar information obtained from wall mounted Langmuir Probes. The neutral pressure and neutral flux density measuremnts in the main chamber indicate a linear trend with αt , whereas for the measurements in the divertor a different dependence on qcyl/Btor is necessary. At the same time, the density shoulder amplitude assessed by a SOL density qualifier saturates when αt exceeds ≈ 0.55 within this database. The e-folding density decay length λn describing the density behaviour in the near-SOL reveals a similar trend. An analysis of filament properties measured by exploiting dedicated probe-heads near the first wall show a comparable linear trend of filamentary convective contribution to heat and particle fluxes with αt. At the same time, it has been identified that the cause of the enhanced filamentary transport in H-mode for QCE scenarios is an increase in particle flux, primarily caused by an increase of filament density and the packing fraction. It is noteworthy that the filament temperature for ions and electrons remain consistently high without substantial variation whenever the high density regime has been achieved. A power balance based on probe and calorimetry measurements suggests that the filament heat flux contribution remains constant, mostly below 25 %, throughout the entire database. The estimation of the gross erosion on the first wall indicates that the erosion is carried by the background contributions and not by filaments. Significant erosion is caused only in the presence of impurities in the far-SOL. Furthermore, the gross erosion appears to saturate when αt exceeds ≈ 0.55 within this database.