The National Fusion Collaboratory Project gave a technology demonstration throughout the week at the SuperComputing 2003 meeting in Phoenix, Arizona as well as a summary oral presentation in the DOE SciDAC booth. The demonstration consisted of the SC03 show floor representing an off-site scientific group collaborating in real time with experimental operation of the DIII-D tokamak. The technology demonstration at this show will be used in tokamak operations in early CY04.
The demonstration consisted of:
The Access Grid was used for audio, video, and shared applications that connected Phoenix and San Diego as well as PPPL. Audio included the DIII-D Chief Operator voice for pulse count down.
Real time pulse waveforms available seconds after the pulse ended via a web browser.
MDSplus for data acquisition as well as for despatching analysis codes.
Shot cycle status and Data Analysis Monitoring (DAM) via a web interface.
Advanced reservation computation to support tokamak operations; between pulse EFIT running on a PPPL cluster yielding 200 time slices in less than 190 seconds.
Between pulse TRANSP with status information available via the Fusion Grid Monitoring System with integrated graphics via ElVis.
X application sharing (many to one) within the SC03 show floor for enhanced group discussion capability.
The EFIT code was recently extended to include the new Li beam edge current data and a new internal ECE/Thomson topological constraint in the equilibrium reconstruction. The ECE/Thomson constraint can be used with or without MSE data to provide an additional condition on the current profile and to improve overlapping of electron temperatures from ECE and Thomson measurements. The response matrix was expanded to also include the Li beam data, with full corrections to the signals due to the viewing geometry. Reconstructions of the edge current profiles using the new Li beam data for DIII-D H-mode discharges show a large edge current component in the pedestal region as expected, whereas the L-mode cases show very little edge current.
A post APS workshop on Error Magnetic Fields, organized by Lang Lao, of General Atomics, was held at the Hyatt Hotel in Albuquerque, NM on October 31 and November 1 2003. Thirty participants attended and 14 presentations were delivered in the day and a half. The PDF files of the Error Field Workshop presentations can be found at http://web.gat.com/~lao/ef/efws2003/
Incorporation of a model for mass shedding due to a differential drift of a pellet cloud has now brought pellet drift modeling and experimental results into closer agreement. Magnetic shear deforms the cloud into an ellipse as it extends along the flux tube, causing a differential poloidal drift of the ends of the cloudlet. Thus different segments of the cloudlet drift out of the electrostatic region so that the end cells are sequentially disconnected and shed. This significantly decreases fuel penetration. The Pellet Relaxation Lagrangian (PRL) code has been upgraded to include this effect and yields good agreement with inside launch pellets experimental results from DIII-D. Refined calculations for ITER now indicate good penetration of the pellet mass for High Field Side injection; without mass shedding, the penetration was predicted to be too deep. These results were reported at the recent APS meeting in Albuquerque (Poster #QP1-2).
These highlights are reports of research work in progress and are accordingly subject to change or modification