September 17-18, 2002
Hyatt Regency Islandia
San Diego, CA

Overview and Orientation

The purpose of the initiative is to make a significant advance toward the ultimate objective of fusion simulation, that is to predict in detail the behavior of any discharge in a toroidal magnetic fusion device on all important time and space scales. We envisage the initiative to be structured so as to add capability incrementally, and that it will adopt near and intermediate term objectives of: supporting basic theoretical research; supporting the understanding, interpretation, and planning of ongoing fusion experiments; and the exploration of new confinement concepts.

The FSP is seen as comprised of three interacting types of activities:

• Fundamentals: basic research in physics, applied mathematics, and computer science needed as the basis for the element modules.
• Applications modules: development of "stand-alone" suites of codes, that address distinct fundamental problem areas such as MHD, turbulent transport, or external sources. Each applications module must be developed to represent the state-of-the-art in physics content, numerical methods, and computational science methods enabling efficient incorporation into the integration framework.
• Integration: development necessary to create a comprehensive simulation capacity.

Further, the integration per se is expected to proceed in stages:

• Incorporation of individual modules into the integration frameworks: The intent is that researchers at the fundamental level and developers of individual applications modules must benefit from integration through such services as easy access to experimental data, advanced mathematical algorithms, and visualization.
• Incorporation of a small number of interoperable applications modules into the computational infrastructure to address critical fusion research issues at the earliest possible time: We expect to develop a small number of Focused Integration Initiatives in such areas as:

  –	Plasma edge/ pedestal physics
  –	Self consistent evolution of pressure profiles with turbulent transport
  –	Island growth, sawtooth oscillations, neoclassical tearing modes
  –	Transport modeling with simplified physics

  It is at this level that approaches are developed to the fundamental integration issues of highly disparate time and space scales, and coupling of models with different dimensionalities.

• Cross coupling of the techniques developed in the previous levels to provide an increasingly comprehensive simulation capability.

Draft "Planning Matrix"

The "planning matrix" is a working document meant to help organize thinking about the meeting and the project. In it the rows — the 'Focused Integration Initiatives' (FIIs) — are meant to be beginnings of integration, i.e., early integrated activities / research problems that people can get their minds around without becoming bogged down by too much generality. Each such FII could last about 2.5 years of a five-year initiative, and a few of the FIIs would then (hopefully naturally) evolve into larger chunks of integration, at about year 3. Such a structure could be a very straightforward way to introduce, among other things, a few different models for interpretative tools and for algorithms.

Notes on the meeting schedule and agenda

1. Sept 17 talks on fundamental application areas — The first goal of these talks should be to familiarize the applied mathematicians and computer scientists with the scientific issues to be addressed in the project and to assist their engagement in the planning and proposal process. To this end, the talks should describe the physics issues and existing physics codes in terms understandable to non-physicists. They should also describe the relevant equations, algorithms and meshes, inputs and outputs; and identify outstanding mathematical and computational problems. A second goal will be to indicate how each area could contribute to and connect to the various integration initiatives. Therefore, the talks should point toward the next steps to be taken in integrating the area into a comprehensive simulation structure, and also to describe unresolved issues of basic physics that must be addressed.
2. Cross cutting sessions on Sept. 17 and FII summaries on Sept. 18 — The goal of these sessions is to identify a few potential Focused Integration Initiatives, which will constitute the near to intermediate term cross-discipline integration projects. These initiatives should address specific important scientific priorities for the fusion program. They should be of a nature to benefit from coupling of a small number of the fundamental discipline areas. And they should serve as test beds for developing approaches to generic integration issues, such as disparate time or space scales, extreme plasma anisotropy, or coupling of models having different natural dimensionality. To structure the process of defining such initiatives, we have identified three broad areas as initial starting points for discussion (see the corresponding rows of the planning matrix) — Plasma edge, Profile evolution, and Island growth. These topical areas should be considered suggestive, not prescriptive. The objective is to develop a compelling case for a few examples. The final decision as to exactly what integration initiatives to include in the project will occur during the proposal stage or perhaps during the project itself.

   The output of each of the Sept. 17 Focused Integration Initiative cross cutting sessions should be:

   • One, or a few, 'over-arching' scientific questions that would satisfy the criteria of importance to the fusion program, along with a concise explanation of this importance
   • Identification of the needs for cross-disciplinary integration
   • Statement of potential to address generic integration issues and possible approaches to these issues
   • Vision for further linkages, perhaps ultimately leading to an approach to a comprehensive model

   A fourth cross-cutting session is planned on Whole Discharge Modeling (bottom row of matrix). The objective of this session is to develop one or more visions of development paths leading from where we are now to the ultimate goal of a comprehensive, integrated predictive and interpretive simulation capability. Any such integration development path must satisfy a number of possibly competing criteria. It must be capable of incremental evolution, providing research benefits in the near to intermediate term. It must be flexible and incorporate approaches to the generic integration issues discussed in the previous paragraph as such approaches are developed in the Focused Integration Initiatives. It must be broad, supporting understanding, development, and optimization of advanced tokamaks as well as stellarators and other, non-tokamak, configurations. A more complete discussion of the integration requirements is contained in section III.C of the July 12, 2002, ISOFS subcommittee report.

   The output of this session should be:

   • One, or a few, visions of such development paths
   • Approaches to maintaining flexibility throughout development
   • Requirements of interpretive modeling and approaches to simultaneous development of predictive and interpretive modeling capability
   • Approaches to including 3D capability from the project outset, without incurring computational penalty to simulation of axisymmetric systems

   In advance of the workshop, the cross cutting session organizers should arrange for appropriate speakers and participants from the fusion and applied mathematics communities in order to bring out the ideas needed for the desired session outputs. The talks should only address the desired outputs of the session and therefore should emphasize ideas and visions for the future. Also, the associated mathematical and computational needs should be discussed such as mathematical structures, associated interpretation tools, and the associated validation issues. Traditional technical talks of recent results are not appropriate. Talks should be limited to about 10 minutes and about 3 slides. The session organizers will also present summary talks on Sept. 18 that should describe the session outputs.

3. Cross-cutting sessions on Sept. 18 — These will be led by the Sept. 18 AM overview speakers; and, physics participants for interpretation in session (a) would participate in the architectures part of session (b), and physics participants for architectures in session (a) would then participate in the interpretations part of session (b).