A set of six NIMROD simulations of the rapid shutdown of diverted DIII-D discharges by argon pellet injection was performed. Three of the experimental discharges produced runaway electron (RE) current plateaus and three did not, and the aim of the simulation was to understand how variations in the initial current profile contributed to this result. The NIMROD simulations use a test-particle model to calculate the fraction of the RE seed that remains confined after the disruption related MHD. With one exception, the NIMROD calculated confined RE fraction was larger when the experimental RE current was also larger across discharges. For the four discharges with q0 < 1, a correlation was found between the fluctuating field amplitude, and therefore the RE confinement, with the initial q95. With only two discharges having q0 > 1, further analysis is needed to determine what features of the current profile affect RE confinement in those cases.
The calculation of electron transport fluxes is incorporated into a numerical code for neoclassical transport that treats the exact Fokker-Planck operator by an expansion in Legendre-Laguerre polynomials. This represents an improvement over existing calculations that employ model collision operators. A new formulation for the electron fluxes is used that preserves the correct poloidal angle dependence of the bootstrap current in general magnetic geometry, in contrast with most existing calculations where the bootstrap current is in a flux-averaged form. This, more accurate, characterization has implications for magnetic field diffusion over long time scales.
In collaboration with W. Guttenfelder of PPPL, nonlinear GYRO simulations of NSTX discharge 120968 at mid radius have found the dominant mode and the resulting electromagnetic transport to be from long-sought micro-tearing modes. The results will appear in the May 2011 issue of Phys. Rev. Lett. The simulated magnetic flutter electron energy transport is comparable to experimental levels, although it is significantly reduced by ExB shear. These first nonlinear simulations are very challenging since the micro-tearing modes require extremely fine radial scale grid resolution. Linear GYRO simulations of the NSTX parameter space have been mapped into regions dominated by ITG, TEM, ETG, and now micro-tearing modes. In an effort to elucidate the observed high L-mode edge transport, similar calculations covering the edge region using expensive fine radial scale GYRO simulations are continuing but are so far unsuccessful in finding micro-tearing modes that explain the missing near L-mode edge transport.
The M3D-C1 code has been modified to allow time-independent calculations of the linear non-axisymmetric plasma response. This method is significantly less computationally expensive than the previously used method, in which the time-independent linear response was calculated by evolving the dynamical system to a steady state. Both methods yield essentially the same result; however, for some cases of practical interest, runtimes are reduced by more than a factor of ten with the new method. Calculations of two-fluid linear response obtained using this method were presented at the Sherwood fusion theory conference in Austin, TX.
Disclaimer
These highlights are reports of research work in progress and are accordingly subject to change or modification