The OMFIT framework has been updated to handle I/O of the BOUT++ files. The BOUT++ code is a modular framework designed for simulating plasma fluid models in curvilinear geometry. A new BOUT++ module for OMFIT is being developed. The new module was used to simulate the evolution of an unstable tearing mode in a slab geometry and the results have been verified against analytic predictions of the tearing mode growth rate. Simulations were executed both on GA workstations and on NERSC supercomputers. In the future, the BOUT++ module in OMFIT will be expanded to investigate ELMs and filament dynamics in the edge of the DIII-D plasma.
The local version of the MARS-F code was placed under version control and is now available for collaborators. Several recent modifications were merged, including an updated set of in-house FFT routines optimized for use in MARS-F. The implementation of version control will aid future code maintenance and upgrades.
A script for computing multiple eigenmodes using the GATO code was developed and appears to be working efficiently. For each eigenmode calculation, the initial eigenvalue guess is set from the history of the previous few eigenvalues so that a minimum number of iterations are required to converge to the desired one. Since stable kink modes comprise the plasma response to external non-axisymmetric fields, it is important to delineate the spectrum. However, in the past, searching for particular modes was always problematic since the stable spectrum contains Alfven and acoustic continua as well as gaps and isolated stabilized kink modes. The new script is also useful for TAE mode studies as well as for following unstable kink modes into the stable spectrum as a parameter such as beta is varied. In the first full set of runs to compute the stable eigenmodes for the DIII-D low q95 discharge #150513 with the continua removed, the expected stable m =2 and m = 3 external kinks were found. The modes, however, were interspersed in a sequence of Sturmian m = 1 modes, and coupled to them as well as to each other, and it is not yet clear if this coupling remains intact in the full inertia case. The full inertia case is being run but requires the calculation of several thousand eigenmodes from the numerically discretized continua in contrast to the approximately fifty eigenvalues present when the continua are removed.
The rCYCLO code has a simple test model ignoring parallel motion and variation so that 3D gyrokinetics can be tested against 4D cyclokinetics and contains high frequency ion cyclotron modes. 4D cyclokinetics dynamically follows the gyro phase in lieu of the gyro phase angle averaging in gyrokinetics. The project aims to quantify the breakdown of gyrokinetics at high turbulence levels (see highlight from July 19 2013 at Theory Weekly Highlights for July 2013). rCYCLO has recently replaced the near adiabatic electron model with a collisional fluid electron model more appropriate to the high turbulence cold L-mode edge where we hope to find that cyclokinetics has a higher transport level than gyrokinetics.
Zhao Deng, a graduate student at Peking University, has arrived for a long term visit with the GA Theory Group to work with Ron Waltz on the reduced cyclokinetics code rCYCLO.
These highlights are reports of research work in progress and are accordingly subject to change or modification