Divertor heat loads due to edge localized modes in ASDEX Upgrade and JET

Transient power loads due to ELMs are a major concern for the life time of divertor tiles in ITER, irrespective of the choice of plasma facing material. Large ELMs cause enhanced material erosion (ablation, evaporation) possibly leading to unacceptable impurity influx, radiation and fuel dilution in the plasma core. Significant progress has been made towards a physics-based model describing Type-I ELM energy transport to the divertor and the first wall. Specifically, the temporal evolution of infra-red measured divertor heat loads due to Type-I ELMs on ASDEX Upgrade (AUG) is well reproduced by an analytic formula representing the arrival of ions originating at the outer midplane with a Maxwellian velocity distribution characterized by only the pedestal. For large Type-I ELMs in JET (100-600 kJ, 5-20 % pedestal energy) and ASDEX Upgrade (6-20 kJ, 3-10 % pedestal energy) the ELM power deposition time on the inner/outer divertor targets is dependent only on the corresponding parallel ion transit time. In particular, for low pedestal collisionality (as expected in ITER), it predicts ELM power deposition time of 1.2 times the parallel transport time which yields a power deposition time of 280 μ s at the ITER pedestal temperature. The analysis of JET and ASDEX Upgrade data also reveals the temporal hierarchy of ELM energy release, parallel transport and power deposition time. On JET, the fraction of energy of deposited on the target within the duration of the power deposition time varies between 20% for the largest ELMs (lowest pedestal collisionality) and to 35% for smallest ELMs (highest pedestal collisionality). It is noteworthy in that respect that the found temporal hierarchy of ELM energy release, parallel transport and power deposition time is broken for the much smaller (and more collisional) Type- III ELMs. For these ELMs, both quantities ELM energy release time and parallel transport have to be taken into account when estimating the resulting ELM peak heat fluxes. The above results confirm the previous, purely empirical scaling, on the type-I ELM power deposition times. However, combined with recent material test, the formerly calculated limit of a maximum ELM energy loss in ITER is revised and a significantly lower value of about 1-2 MJ is found. Therefore, some form of ELM mitigation techniques appears to be mandatory for ITER. (author)