Spatio-temporal investigations on the triggering of pellet induced ELMs

Abstract

Pellets injected into type-I ELMy H-mode discharges are known to trigger edge-localized modes (ELMs). In order to understand the underlying processes the triggering mechanism was investigated in this paper. The major questions of the investigations to be answered were: at which magnetic surface was the ELM initiated and what was the corresponding perturbation caused by the ablating pellet? During the investigations the natural ELM cycle was probed by injecting pellets from the high field side of the ASDEX Upgrade tokamak with significantly lower frequency than the natural ELM frequency. To determine the location of the seed perturbation of the ablating pellet triggering an ELM, the dynamics of the triggered ELMs was linked to the time history of the pellet position in the plasma. The ELM onset was determined by analysing magnetic pick-up coil signals and its delay relative to the time when the pellet crossed the separatrix was measured as a function of the pellet velocity. Supposing that to trigger an ELM a pellet has to reach a certain magnetic surface of the plasma independently of its mass and velocity, the most probable location of the seed perturbation was found to be at the middle of the pedestal—at the high plasma pressure gradient region. The onset of the MHD signature of the ELMs was detected about 50 µs after the pellet reached the seed position. According to our observations ELMs can be triggered either by the cooling of the pedestal region causing sudden increase of the pedestal plasma pressure gradient driving the plasma to the unstable region of the ballooning mode or by the strong MHD perturbation triggering an instability developing into an ELM.

Investigation of pellet-triggered MHD events in ASDEX Upgrade and JET

“…The magnetic perturbation decays quickly (within about 50µs -which is about the same time needed to trigger an ELM after the corresponding perturbation has been imposed [15]) once the driving pellet ablation vanishes after pellet burn out. Hence, it is quite obvious the magnetic perturbation is driven purely by the pellet and no growing mode activity was triggered.…”

“…On account of the practical importance, pellet triggering investigations in the type-I ELM regime are so far the most advanced ones, so that this study can be regarded as an extension of these previous investigations. We therefore recapitulate here just their main results, described in more detail in [15] to put them into the context of a broader view on the pellet triggering of ELMs or MHD activity in general.…”

“…Specific AUG investigations [15] of the nature and dynamics of the trigger process showed that the pellet-imposed local perturbation triggering the ELM onset is imposed within the edge transport barrier but clearly inside the separatrix. A delay is observed between the implementation of the initial local perturbation and the ELM onset of about 50 µs -a value coinciding, as already mentioned, with the observed decay time of a purely pellet driven MHD perturbation in OH phases.…”

ELM pacing and high-density operation using pellet injection in the ASDEX Upgrade all-metal-wall tokamak

“…Dependence on the pellet speed and mass was also observed, but it remained unclear if this was due to corresponding variations of the resulting pellet perturbation or to the limited technical performance of the high-frequency pellet injector (HFPI) [23] falling short of delivering all pellets in viable conditions. Pellets are regarded as viable for ELM triggering when providing sufficient mass and speed allowing penetration to the pressure profile pedestal top [24,25]. Primarily, HFPI was built with its performance designed to meet all viability requirements for ELM control and mitigation after the ELM impact had been found to potentially damage the JET divertor [26].…”

Comparison of pellet acceleration model results to experimentally observed penetration depths

“…The centrifuge injector system [4], capable of delivering pellets in three different sizes (1.6, 2.7, 4.0 ×10 20 D atoms), at several velocities (240, 600, 880 and 1000 m/s) and at frequencies up to 83 Hz, was used. Several plasma scenarios with pellet injection were selected to benchmark the codes: type-I ELMy H-mode discharges with smallest size pellets and several pellet velocities [5], as well as other types of discharges with lower heating power and/or large pellet size to achieve deep penetration (see Table 1 for details). Table 1.…”