An improved simulation for an NSTX high harmonic fast wave (HHFW) heating discharge using the coupled ORBIT-RF and AORSA codes (see highlight October 16 2009 at Theory Weekly Highlights for October 2009) found better quantitative agreement between the simulation fast ion spatial profile and the fast ion D-alpha (FIDA) spectroscopic data. Parallel energy diffusion and energy scattering effects that had been ignored in the previous simulations were implemented using a more general expression for the fast-ion energy slowing down process. A sensitivity study with respect to the convergence and particle statistics was performed with an improved Coulomb collision model. The simulation was also extended three times longer than previously and iteration between the fast ion distribution and the RF wave fields was repeated to obtain convergence. The sensitivity of the statistics in terms of the timing sequence for extracting fast ion information was also extensively studied. Poor conservation of test particles in the previously computed fast ion data, which caused unrealistic losses of the fast ion population and contributed to the poor comparisons, was improved. Although better agreement with the FIDA data was obtained, the simulation still finds a larger outward radial shift than the FIDA data shows. Further sensitivity study and physics improvement are underway.

A new version of the EPED model, EPED1.6, which includes a new, more comprehensive model of diamagnetic stabilization, has been successfully tested on a set of C-Mod, DIII-D and JET discharges, finding good initial agreement. The recently developed and successfully tested predictive model for the pedestal, EPED, (see highlight for January 8 2010 at Theory Weekly Highlights for January 2010) combines calculated peeling-ballooning and kinetic ballooning mode constraints to provide predictions of the pedestal height and width. The constraints are first principles in that no parameters are fit to observations. Recent comparisons and a dedicated experiment on Alcator C-Mod revealed the importance of highly accurate treatment of diamagnetic stabilization of peeling-ballooning modes, due to the high diamagnetic frequency characteristic of C-Mod. A new diamagnetic stabilization model was therefore developed using simulation results from the BOUT++ code, which includes non-local diamagnetic effects. P.B. Snyder will be visiting MIT in May to work with J. Hughes and the C-Mod team on continued analysis and model testing using results of the previous dedicated experiment, and to plan future experiments to further test the model as part of the 2011 Joint Research Target on pedestal structure.

The effects of counter ECCD on tearing mode stability and saturated magnetic island width were evaluated using a series of kinetic EFIT equilibria reconstructed from a DIII-D counter-ECCD tearing mode experiment. The reconstruction used MSE and kinetic profile data with a local current model to represent the counter ECCD. The resistive stability properties of these equilibria were analyzed using the resistive MHD code PEST3 and the Lehigh University NTCC island width module ISLAND. Initial results indicate that the tearing stability index and the saturated m/n = 3/2 island width increase strongly with a narrow ECCD width around the q = 3/2 surface, consistent with experimental observations. Work is underway to estimate the saturated 3/2 island width using the Modified Rutherford equation coupled with PEST3 and the 3D MHD code NIMROD.

An analytical model for the decay of runaway electrons (REs) after a massive gas injection found that the RE current decays a factor of two after 30 ms following a sudden eleven-fold increase in total electron density and five-fold increase in electric field, consistent with preliminary data from the recent neon gas injection experiments on DIII-D. The model was developed to determine whether “mature” high-current REs, once formed, could be stopped after a massive gas attack. The important model assumption is that mature runaways have the energy distribution function that is expected if the avalanche production mechanism were responsible for their formation. Using the characteristic equation for the relativistic particle momentum trajectory, and given that pitch angle scattering for highly relativistic particles is much weaker than drag, the time evolving energy distribution function is then solved for times after an assumed instantaneous background particle density increase in the RE channel region. The unknown internal electric field is obtained by including self-consistently the loop voltage increase associated with the calculated current decay rate, assuming a fixed RE channel inductance. The result lends support to the avalanche mechanism as the cause of the long-lived runaway plateau in DIII-D often seen during disruption mitigations.

The NEO code has been used to analyze the deuterium ion flow of DIII-D L-mode plasmas in the edge region. To enable these studies, a new method was developed, based on using the carbon toroidal flow measurements at the outboard midplane as a calibration, to determine the shift in the parallel flow due to the radial electric field Er. This method is useful when the poloidal flow measurements are either not available or have large uncertainties. So far the method has been implemented only in the weak rotation limit, since in that case the neoclassical flow coefficient does not depend on Er. Using this method, the NEO results for the deuterium flow, which were calculated inside the separatrix, and the measurements from outside the separatrix, approach each other asymptotically indicating that the flow remains essentially neoclassically driven even close to the plasma edge.