Evaluation of the Porcelli sawtooth crash criteria through a complete giant sawtooth cycle was completed for the neutral beam and fast wave heated discharge #96043, and shows complete consistency with the observed crash. While the Porcelli model using simplified expressions for the key contributions has been found to predict average sawtooth periods reasonably well in many cases, it is not completely adequate for quantitatively reproducing the actual crash time for a specific sawtooth in an experiment. The new calculations used the GATO code for the ideal contribution to the crash criteria for a series of six successive time slices during the sawtooth cycle, as described in the highlight for April 06 2007, and the ORBIT-RF and TORIC codes to accurately calculate the non-Maxwellian fast-ion pressure and its kinetic contribution. Within the estimated uncertainties in the equilibrium reconstruction and fast ion distribution modeling, the calculations predicted the trigger criterion becomes marginally satisfied right before the observed sawtooth crash. Future work will focus on reducing key uncertainties and extending the calculations to normal sawteeth as well as other giant sawtooth discharges.
The DIII-D computational resource-monitoring infrastructure is now up and running. This uses the open source Ganglia monitoring system and reports real-time information on the status of DIII-D-related computational servers. The reports include a variety of performance metrics, such as memory, CPU load, disk usage, and MDSplus usage, as well as network statistics. Currently, the system is monitoring the STAR, LSF, and LOHAN clusters, the energy group web server, and the ATLAS server (total 30 servers). The new system is vital for monitoring the overall health of the computing resources being used at DIII-D. It also provides information about usage patterns, which helps inform decisions on hardware upgrades. The real-time monitoring report is located at: http://anubis.gat.com
A closed form solution for total number of impurity ions assimilated in the plasma, Ni, during massive gas jet injection (MGI) for disruption mitigation was found in terms of the vacuum fueling rate and cross-field diffusion coefficient D(t) of the impurity ions. From interferometer measurements of line electron density, which give Ni, we can then infer the diffusion coefficient at any time during the thermal collapse. Normally, the jet does not penetrate the plasma directly, but rather the neutral gas density builds up in the vacuum region, ionizes in an ultra thin layer at the plasma surface, and the ions are drawn inward by some sort of diffusion process during the thermal quench. Conservation of particles leads to a Volterra integral equation for Ni that can be solved formally by a Laplace transform. For the high flow helium Medusa jet on DIII-D, this yields a large diffusion coefficient D = 23 m2/s; typically, in normal tokamak operation without MGI, plasma species diffusion coefficients are ~ 1 m2/s.
GYRO simulations with kinetic electrons have verified the existence of pinch effects in the radial flow of toroidal angular momentum. A momentum pinch is needed to explain the spontaneous toroidal rotation in C-Mod and DIII-D discharges in the absence of any external torque applied to the core. Spontaneous toroidal rotation during heating without an external torque was shown to follow from the off diagonal nature of the toroidal viscous stress in earlier work by Staebler in 2001. In work presented at the recent TTF and Sherwood meetings, GYRO has verified both the E x B shear pinch effect predicted by Dominguez and Staebler in 1993 and the more recently discovered pinch effect from curvature drift with finite parallel velocity. The momentum pinch effects will be built into the TGLF transport model.
These highlights are reports of research work in progress and are accordingly subject to change or modification