Analysis of the periodic variation of pellet ablation radiation intensity in ASDEX Upgrade

Abstract

In a future fusion reactor, the main fuelling method will likely rely on the injection of solid hydrogen pellets. Current predictions assume that this goal can be achieved, since being based on a technology which is already largely developed. However, this belief is founded on modelling tools that are usually aligned to the observation made in existing devices and then extrapolated to reactor conditions. This approach needs a sound consideration of its intrinsic restrictions and any observed feature not reproduced by the utilised codes should be applied to check their validation and possibly contribute to their refinement. One specific feature still lacking an explanation of a reasonable and self-consistent mechanism in the current models is the appearance of a phenomenon called striations, which are high frequency variations in the radiation emitted during the pellet ablation process. In order to provide a sound and reliable database for further considerations, a dedicated analysis of this effect has been performed on the mid-size tokamak ASDEX Upgrade. Therefore, such cases have been selected with the relevant signal recorded with sufficient temporal resolution during experiments covering a wide variation of plasma and pellet parameters which are regarded to be potentially influential on the striation pattern. In addition, it was ensured that for any specific case the observed behaviour was reproducible for several individual ablation events under identical conditions. In all cases considered, the observed radiation-intensity variations appear with a typical pattern showing a broad peak of frequencies in the range 50–150 kHz. This characteristic unveils a notable resilience against any parameter variation. This new collection of data can now act as firm basis to corroborate future modelling code-validation efforts. In addition, the analysis method can provide a relatively simple way of reviewing future modelling predictions.