In an effort to validate the predictive capability of extended-magnetohydrodynamic simulations of disruption mitigation for future devices, the M3D-C1 code is being used to model JET plasmas with realistic, shattered-pellet-injection (SPI) plumes. We consider two different 8.1-mm cylindrical pellets used in the JET experiments: a pure-neon pellet traveling at 150 m/s and a 95%/5% deuterium/neon pellet traveling at 300 m/s. A statistical fracture-threshold model is used to divide these pellets into a collection of smaller fragments. A cloud of uniform fragments is then constructed such that it has the same average ablation rate as the statistically computed cloud (at constant electron density and temperature). For pure-neon, this results in 30 1.71-mm fragments. For the mixed pellet, the cloud has 85 1.21-mm fragments. The fragments are then distributed longitudinally and angularly to create a realistic cone to be injected into a high-thermal-energy (Scenario 1) JET equilibrium. M3D-C1 modeling shows that both plumes result in similar radiation peaks and amplitudes of MHD activity, though the mixed pellet results in a higher electron density due to the abundance of deuterium. In addition, the mixed pellet quenches the plasma faster, but this is due to its having twice the speed of the pure-neon pellet. The mixed pellet propagates much farther into the plasma before the onset of MHD instability and generally shows a much more quiescent quench as the fragments propagate into the plasma core. In the near-term, the radiation fractions of these simulations will be validated against JET data. Numerical studies with M3D-C1 will also be performed to determine the separate effects of pellet composition, shard size, and pellet velocity.

Single seeded blob simulations using parameters of a DIII-D SOL have been carried out with the gyrokinetic code Gkeyll to explore the effect of neutral interactions on radial blob velocities. Theoretical velocity scalings were adapted to account for finite ion temperature in the gyrokinetic simulations, and Gkeyll results demonstrate reasonable agreement. Simulated blob velocities are comparable to measured DIII-D experimental values (Boedo, J.A. et al. Phys. Plasmas, 2001). In simulations with static background neutrals and only the charge exchange interaction, increasing the background neutral density reduced blob velocities, in accordance with theoretical predictions and experimental observations (Katz, N. et al. Phys. Rev. Lett., 2008).