In the first comparison between the measured lifetime of small lithium pellets (granules) injected into DIII-D and the theoretically predicted lifetime based on a new ablation model, agreement between theory and experiment for the largest sized pellets was excellent but an increasing mismatch was found for smaller pellets. Single granules were injected with velocities of 109 m/s into two different discharges with respectively high and low torque in order to gain an understanding of the ELM triggering process. The pellet penetration model employed Osborne functions to model the steep temperature and density profiles in the proximate pedestal region. The granules injected were in the range from 200 to 900 microns, small enough to ensure that most of the pellet material is completely deposited inside the pedestal region. For the smallest pellets, the predicted lifetime was about a factor two too long. A possible explanation proposed for the discrepancy at smaller pellet size may be that ion ablation was not included; the ablation cloud of a smaller pellet is thinner and more transparent to incident plasma ions, which carry additional energy flux to the pellet surface. For typically sized cryogenic fueling pellets, the plasma ions do not take part in the ablation process.

A report entitled “TGLF Recalibration for ITER Standard Case Parameters” for the 2015 DoE Theory and Simulation Performance Target was authored by the members of the GA theory group. The work used on the order of 100 nonlinear GYRO simulations, for ITER-relevant parameters, to produce the dataset required for improvement of TGLF under reactor conditions, in particular, for near threshold transport in the absence of driven plasma rotation. Motivated by this data, a new saturation rule was developed for TGLF in which the shearing effect of nonlinearly generated zonal flows can be incorporated into the model. By considering additional multiscale simulations, not connected with the milestone, the new saturation rule was also generalized to more accurately include the high-k contribution from ETG to the electron energy transport. While this new TGLF capability needs further refinement - in particular for the high-k component - we believe it represents a significant advance in modeling capability.

A new project has been initiated to formulate a code for 6D full kinetic simulation of drift wave turbulence by expanding the velocity space distribution function in Gaussian Radial Basis Functions (GRBFs). Essentially the distribution function at each space point is represented by many mini-Maxwellians centered on 3D velocity vector “grids”. This representation can be used to exactly represent the nonlinear inverse-square force Fokker-Planck collision operator. (J. Candy, APS-DPP 2015 invited talk). The collisionless side of the 6D full kinetic Vlasov equation takes a novel form in which there are no velocity gradient operators on the velocity grid and only spatial gradients in convection terms appear. As a first exploratory problem, a new code (LEDGE6D) was designed to simulate drift wave turbulence, with ion temperature gradient (ITG) modes and adiabatic electrons in tokamak geometry. The single processor version of the new code found good agreement with the rCYCLO code (Zhao Deng, APS-DPP 2014 invited talk) for the high frequency ion cyclotron modes. Benchmarking against the GYRO 5D code for the ITG mode growth rates requires a parallel version of the code, which is currently being developed.

**Disclaimer**

These highlights are reports of research work in progress and are accordingly subject to change or modification