Workshop to Plan Fusion Simulation Project

Transcription

1 Workshop to Plan Fusion Simulation Project (Tokamak Whole Device Modeling) Presented by Arnold H. Kritz Lehigh University Physics Department Bethlehem, PA 18015, USA FESAC March 2, 2007

2 FSP Objective and Motivation Primary objective of FSP Create high-performance software to carry out comprehensive predictive integrated modeling simulations, with high physics fidelity, relevant to ITER and other tokamaks Leadership class computers will be necessary to achieve this objective Urgent need for FSP motivated by fact that there are significant physics problems associated with ITER discharge scenario planning and control Prior to completion of ITER construction, controls must be developed to suppress large scale instabilities that can adversely affect confinement Accurate predictions are needed for Edge transport barrier that enhances the core plasma confinement Edge instabilities that cause fluctuations in power to the divertor and first wall Each discharge in ITER is expected to cost about a million dollars Whole device computer simulations needed to optimize discharge scenarios Fully verified and validated comprehensive integrated modeling capability is essential to support ITER as well as other current and planned tokamaks FESAC March 2, 2007

3 FESAC March 2, 2007 FSP Background Previous Fusion Simulation Project (FSP) to address questions outlined in charge letter, February 22, 2002 Develop 5-6 year initiative the goal of which is develop an improved capacity for Integrated Simulation and Optimization of Fusion Systems FESAC appointed committee to develop a roadmap Final FESAC Report December Fusion Simulation Project (FSP) envisioned as a 15 year, $20M/year multi-institutional project Develop a comprehensive simulation capability for magnetic fusion experiments with a focus on ITER Recommended an approach through Focused Integration Initiatives Coupling pairs of components before moving to whole device modeling OFES formed an FSP Steering Committee in 2003 Develop project vision, governance concept, and roadmap for the FSP Recommends that the FSP consist of three elements: Production component, a research and integration component, and a software infrastructure component Final report in Journal of Fusion Energy, Vol. 23, No. 1, March 2004

4 FESAC March 2, 2007 Context for FSP It is clear that computer plasma physics simulations have become significantly more sophisticated in recent decades The first round of SciDAC projects concentrated on first principles simulations of individual physical phenomena Development of high fidelity physics models for individual physical processes needs to continue The second round of SciDAC projects combined pairs of processes FSP is intended, after 15 years to combine all relevant physical phenomena in comprehensive whole device simulations In the first 5 years, there will be focus on a limited number of problems for which advanced simulation capability can provide exciting scientific deliverables that substantially impact realistic predictive capabilities Computer hardware is evolving beyond all expectations Within the next five years, we will be in age of petascale computing (10 15 FLOPS) with massively parallel computers Integrated modeling is particularly challenging because of the diverse physics and algorithmic modules FSP will develop comprehensive whole-device modeling With simultaneous interactions of multiple physical processes treated in a self-consistent manner Using modules with much improved physics fidelity

5 Elements of an Integrated Tokamak Modeling Code Sawtooth Region (q < 1) Core Confinement Region Magnetic Islands Edge Pedestal Region Scrape-off Layer Vacuum/Wall/ Conductors/Antenna Core Transport Edge Transport Plasma Turbulence MHD Equilibrium Plasma-Wall Interactions Large Scale Instabilities Radiative Transport Atomic Physics Heating Current Drive

6 FESAC March 2, 2007 Physics is Interactive Many physical processes in tokamaks interact strongly Whole device integrated modeling codes are needed to simulate strongly interacting physical processes observed in experiments Examples of interacting processes: Large scale instabilities can strongly modify plasma profiles and are strongly affected by the plasma profiles Sawtooth oscillations that are triggered by kink or m=1 tearing modes redistribute current density, thermal particles and fast particle species Neoclassical tearing modes (NTMs) are very sensitive to current and pressure profiles and produce flat spots in those profiles Boundary conditions strongly affect core plasma profiles H-mode pedestal height, normally limited by ELM crashes, controls core temperature profiles since anomalous transport models are stiff Wall conditioning has a strong effect on hot ion scenarios Distortion of velocity distribution due to slowing down of fast ions from NBI, RF and fusion reactions need to be included in gyrokinetic turbulence codes Fast ions are redistributed by large scale instabilities and slowing down time is affected by plasma profile changes caused by sawtooth crashes

7 Questions and Plan for FSP Current funding of simulation efforts clearly insufficient to achieve FSP requirements Dramatic increase in funding is required if FSP goals are to be achieved To initiate the FSP project, it is necessary to answer the questions: What are the critical technical issues facing the fusion program? How can high performance computer simulations contribute to the resolution of these issues? What substantial contribution can computer simulation make that traditional theory or experiment, by themselves, cannot? What new contributions can result from advances in physics models, algorithms, software, and computer hardware? What investments in fusion science as well as computational science and infrastructure must be made to obtain the needed answers? How should the Fusion Simulation Project be organized and managed to address these critical technical issues? To obtain required answers, OFES and OASCR have formed a Fusion Simulation Project committee to plan and hold a workshop FESAC March 2, 2007

8 FESAC March 2, 2007 FSP Workshop Committee Co-Chairs: David Keyes (Columbia Univ.) Phil Colella (LBNL) Martin Greenwald (MIT) Dan Meiron (Cal Tech) Arnold Kritz (Lehigh Univ.) Scott Parker (Univ. Colorado) Cynthia Phillips (PPPL) Tom Rognlien (LLNL) Andrew Siegel (Univ. Chicago/ANL) Xianzhu Tang (LANL) Pat Worley (ORNL) Workshop: Wednesday and Thursday May 16-17, 2007 Venue will be in the vicinity of Washington, DC FSP Committee members will continue on Friday, May 18, to complete a final draft of workshop report

9 FESAC March 2, 2007 FSP Workshop May 16-17, 2007 Forty five scientists, representing twenty institutions, from the OFES and OASCR communities will participate on four panels Workshop will include the opportunity for comments by observers FSP Workshop Web site has been established Web discussion boards are in the process of being developed The result of the workshop will be a report containing the conclusions and recommendations of each panel It is anticipated that the result of the work of the FSP panels will be reviewed by a FESAC appointed subcommittee The FSP Workshop will be patterned after DOE-BES Workshops Examples of 17 reports that have been written by Basic Energy Sciences (BES) workshops are available on the Web at

10 FESAC March 2, 2007 Panel Structure Project Structure and Management P. Colella, M. Greenwald, D. Keyes, A. Kritz Integration and Management of Code Components D. Meiron, T. Rognlien, A. Siegel Status of Physics Components S. Parker, C. Phillips, X. Tang Status of Required Computational and Applied Math Tools P. Colella, D. Keyes, P. Worley

11 Project Structure and Management Panel Phil Colella LBNL Martin Greenwald MIT David Keyes Columbia Arnold Kritz Lehigh Panel Issues: Scope of the FSP (ITER requirements) General management structure Establishing project Gantt charts Process when deadlines are not met FESAC March 2, 2007 Don Batchelor ORNL Vincent Chan GA Bruce Cohen LLNL Steve Jardin PPPL David Schissel GA Dalton Schnack Wisconsin Frank Waelbroeck Texas Michael Zarnstorff PPPL Coordination and management of geographically distributed teams Interaction with other related national and international projects How to allow for alternative approaches

12 Integration and Management of Code Components Panel Dan Meiron Cal Tech Tom Rognlien LLNL Andrew Siegel ANL/U. Chicago Panel Issues Component coupling framework issues Project Phasing Duration and sequencing of various phases Validation and Verification procedures Code version control and management Coding standards Michael Aivazis CalTech Rob Armstrong Sandia David Brown LLNL John Cary Tech-X Lang Lao GA Jay Larson ANL Wei-Li Lee PPPL Doug McCune PPPL Mark Shepherd RPI FESAC March 2, 2007

13 Status of Physics Components Panel Scott Parker U. Colorado Cynthia Phillips PPPL Xianzhu Tang LANL Panel Issues: Completeness and robustness of physics modules NBI and RF and nuclear heating Fueling Current drive Transport and turbulence Large-scale instabilities Plasma edge Wall and atomic physics Feedback control Energetic particles What can be achieved in various time frames FESAC March 2, 2007 Glenn Bateman Lehigh Paul Bonoli MIT CS Chang NYU Ron Cohen LLNL Pat Diamond UCSD Guo-Yong Fu PPPL Chris Hegna Wisconsin Dave Humphreys GA

14 Status of Required Computational and Applied Math Tools Phil Colella LBNL David Keyes Columbia Pat Worley ORNL Panel Issues: Numerical algorithms Discretization, adaptivity, solution, optimization Data handling Graphics and visualization Scalability to highest-end platforms Performance evaluation and performance engineering Jeff Candy GA Luis Chacon LANL George Fann ORNL Bill Gropp ANL Chandrika Kamath LLNL Valerio Pascucci LLNL Ravi Samtaney PPPL John Shalf LBNL FESAC March 2, 2007

15 FESAC March 2, 2007 FSP Committee Activities The FSP Committee was convened in January 2007 There have been six conference calls to date Four-panel FSP workshop structure was accepted on 16 January There have been extensive committee discussions about the overall objectives and scope of the Fusion Simulation Project Beginning to focus on the specific deliverables, particularly those that will be delivered at the end of first five years Members for the four FSP panels were chosen by mid-february Careful consideration was given to expertise of panel membership as well as geographical and institutional balance Public FSP web site was established Discussion board and Wiki parts of web site are being established in order to facilitate community input, panel interaction and report writing Conference calls have begun within each panel A scribe will be chosen for each panel to maintain a written record Organization and agenda of Workshop are currently being developed

16 Conclusions The FSP committee and workshop will: Address issues associated with project structure and management of the proposed FSP Identify the critical scientific and technical challenges for the fusion program for which predictive integrated simulation modeling has a unique potential for providing answers in a timely fashion, In a way that traditional theory or experiment by themselves cannot Establish a clear plan to improve the fidelity of the physics modules required for predictive tokamak whole device modeling Identify the critical areas of computational science and infrastructure in which investments would likely produce the tools required for the FSP to achieve its goals It is essential that we produce, in a timely way, advanced simulation capability in support of ITER that can provide exciting scientific deliverables which substantially impact realistic predictive capabilities There is a need to make a clear distinction between new individual scientific discoveries driven by leadership class supercomputers and useful integrated models that are capable of delivering whole device simulations with significantly improved validation FESAC March 2, 2007