Dispersive Shell Pellet Modeling Validation and Extrapolation to ITER*

Dispersive shell pellet (DSP) injection as an alternate disruption mitigation technique to shattered pellet injection (SPI) was demonstrated in DIII-D experiments [Hollmann, Phys. Rev. Lett. 122, 65001 (2019)] wherein diamond shells filled with boron dust triggered an inside-out thermal quench (TQ). MHD modeling of DSP reproduces several experimental trends and suggests that a predictive model is feasible. Simulations explain the generation of runaway electrons only for the most centered payload, which also maximizes TQ mitigation efficiency. Modeling also replicates the experimentally observed reduction of the I_p-spike amplitude for faster pellets. The I_p-spike results from edge current-profile flattening by a 3/1 double tearing mode, whose non-local radial structure stochasticizes much of the plasma volume. Assuming faster pellets, extrapolation of the shell material quantity to ITER should benefit from the reduced surface-to-volume ratio for larger pellets. Initial ITER shell and (beryllium) payload quantity scoping indicates that the payload needed for an ITER TQ is less than what would be obtained using the DIII-D quantity scaled up by the stored thermal energy (roughly 360x).