FPN02-60

Fusion Simulation Initiative Progress

September 13, 2002

In February 2002, the DOE Office of Science asked the Fusion Energy Sciences Advisory Committee (FESAC) to assist in defining a major new initiative to be sponsored jointly by the Office of Fusion Energy Sciences (OFES) and the Office of Advanced Computer Science Research (OASCR). The goal of this initiative, the Fusion Simulation Project (FSP), is to create a comprehensive set of theoretical fusion models, an architecture for bringing together the disparate physics models, combined with the algorithms and computational infrastructure that enables the models to work together. The required funding level is expected to be on the order of $100M spread over five to six years.

A FESAC ISOFS (Integrated Simulation and Optimization of Fusion Systems) Subcommittee, with members from the fusion, applied mathematics and computer science communities, has been constituted to generate a plan for moving forward with the FSP. The FSP undertaking represents a significant opportunity and a significant challenge to the fusion program.

The FESAC ISOFS Subcommittee is hosting a workshop at the Hyatt Islandia in San Diego on Sept. 17-18, 2002, to discuss the goals, framework and governance for the FSP.

At the September 11 FESAC meeting in Gaithersburg, Maryland, ISOFS Subcommittee chair Jill Dahlburg gave a progress report. She said that the Subcommittee would recommend "that a major initiative be undertaken" whose purpose "is to make a significant advance within five years toward the ultimate objective of fusion simulation -- to predict accurately the behavior of plasma discharges in a toroidal magnetic fusion device on all relevant diverse time and space scales." She said, "This is in essence the capability for carrying out 'virtual experiments' of a burning magnetically confined plasma, implying predictive capability over many energy confinement times, faithful representations of the salient physics processes of the plasma, and inclusion of the interactions with the external world (sources, control systems and bounding surfaces)." The progress report states, "Two fundamental issues are common to many fusion physics integration areas: coupling of phenomena at disparate space and time scales, and coupling models of different dimensionality. To solve these generic problems and achieve the integration we are seeking, strong collaboration and advances in physics, applied mathematics, and computer science will be required. This disciplinary integration will be an essential element of the program."

The progress report goes on to say "To succeed, a central feature of this initiative must also be an intensive and continual close coupling between the calculations and experiments. The phenomena in magnetic fusion devices, the equations describing them, and the interactions among the various critical phenomena are sufficiently complex that developing the most effective approximations and establishing when the models have the desired accuracy can only be accomplished by continual iteration and testing of the models against experimental data."

Dahlburg said, "We envision that there will be three major categories of activity within the Fusion Simulation Project: 1) research to advance fundamental capabilities in fusion science, applied mathematics and computer science that address specific program critical needs; 2) development of applications modules, the core building blocks of an integrated modeling capability, in extended MHD, turbulence and transport, the plasma edge, and external sources, including development of the required computational science framework as well as the visualization and interpretation tools; and 3) project integration, which includes development necessary to create a comprehensive simulation including the interconnection and interoperation of multiple applications modules, oversight of software standards and release policies, collaborative tools, program governance, and accountability. It will likely be necessary to support more than one approach to various research areas in all three categories listed above."

Details on this effort can be found at: http://www.isofs.info.

ISOFS Subcommittee Members are: Jill Dahlburg (Subcommittee Chair), Jim Corones (Subcommittee Vice-Chair), Don Batchelor, Marsha Berger, Randall Bramley, Martin Greenwald, Stephen Jardin, Sergei Krasheninnikov, Alan Laub, William Lokke, Jean-Noel Lebouef, John Lindl, Marshall Rosenbluth, David Ross, Dalton Schnack, and Harold Weitzner.