In ITER pellets are envisaged for ELM control and fuelling. More important, ELM control, particularly control of the first ELM needs to be demonstrated already in the non-nuclear phase of ITER during operation in H or He. Whilst D pellets have been established as ELM control technique in the stationary phase with D target plasmas in devices with C as plasma-facing component, the question of other isotopes and non-stationary phases are not so well known. Here, we report on new pellet triggering experiments in ASDEX Upgrade and JET mimicking specific ITER operating scenarios. Both machines are equipped with an all-metal wall, where recent investigations have shown that pellet triggering and pacing become more intricate. In both machines ELM triggering by D pellets injected into D plasmas during extended ELM-free phases, often following the L → H transition, has been demonstrated. In both devices the pellets are found to induce ELMs under conditions far from the stability boundary for type-I ELMs. Near the L → H transition, induced ELMs in some cases might more likely have type-III rather than type-I characteristics. Furthermore, in ASDEX Upgrade this study was conducted during L → H transitions in the current ramp-up phase as envisaged for ITER. In addition, the pellet ELM trigger potential was proven in ASDEX Upgrade with a correct isotopic compilation for the non-nuclear phase in ITER, viz. H pellets into either He or H plasmas. Results from this study are encouraging since they have demonstrated the pellets' potential to provoke ELMs even under conditions quite far from the stability boundaries attributed to the occurrence of spontaneous ELMs. However, with the recent change from carbon to an allmetal plasma-facing components examples have been found in both machines where pellets failed to establish ELM control under conditions where this would be expected and needed. Consequently, a major task of future investigations in this field will be to shed more light on the underlying physics of the pellet ELM triggering process to allow sound predictions for ITER.
Keywords: ELM control, tokamak, pelletExtended 41th EPS contributed paper O3.115 Accepted Version V4.0, 2.3.2015 2
INTRODUCTIONOperating ITER in the reference inductive scenario at the design values I P = 15 MA and Q DT = 10 relies on good H-mode confinement facilitated by the presence of a strong edge transport barrier and a sufficiently high plasma pressure pedestal. The steep gradients evolving at the edge can drive MHD instabilities resulting in an Edge-Localized Mode (ELM) producing a rapid energy burst from the pedestal region. Without dedicated ELM control, the resulting transient heat loads on plasma-facing materials in ITER become critical for operation at a plasma current I P ≈ 9.5 MA [1]; progressing to a higher I P would result in an intolerably short lifetime of the divertor plates [2]. Currently, several options are being considered for this inevitable ELM actuation, but all of them need further validation for the ITER tasks. Obviously...