Public Abstract

DE-SC0024548: HifiStell: High-Fidelity Simulations for Stellarators

Award Status: Active

Institution: The Trustees of Princeton University, Princeton, NJ
UEI: NJ1YPQXQG7U5
DUNS: 002484665

Most Recent Award Date: 07/19/2024
Number of Support Periods: 2
PM: Halfmoon, Michael

Current Budget Period: 09/01/2024 - 08/31/2025
Current Project Period: 09/01/2023 - 08/31/2027
PI: Bhattacharjee, Amitava

Supplement Budget Period: N/A

Public Abstract

HiFiStell: High-Fidelity Simulations of Stellarators Amitava Bhattacharjee, Princeton University David Bindel, Cornell University Scott A. Klasky, Oak Ridge National Laboratory Javier Hernandez Nicolau, University of California – Irvine Elizabeth J. Paul, Columbia University Alkesh Punjabi, Hampton University Aaron Scheinberg, Jubilee Software Development, Inc. Mark Shephard, Rensselaer Polytechnic Institute Carl Sovinec, University of Wisconsin – Madison Ben Zhu, Lawrence Livermore National Laboratory In the stellarator, the confining magnetic field is produced mainly by external current-carrying coils. Since the stellarator can be designed to be current-free, it is intrinsically steady-state, has low recirculating power, and is free of current-driven disruptions, which probably presents the most serious challenge to tokamak operation. On the other hand, stellarators are intrinsically three-dimensional (3D) in that there is no rotational symmetry, which makes the engineering design challenging. Over the last few decades, a new concept has emerged in stellarator design, giving rise to a renaissance: the remarkable discovery that 3D magnetic confinement devices with quasisymmetry can have the same virtues as tokamaks while overcoming some of their inherent drawbacks. The current SciDAC project, HifiStell (High Fidelity Simulations of Stellarators), is a multi-institutional effort consisting of seven Universities, two DOE national laboratories, and a software company, focusing on some key science and engineering challenges pertaining to stellarators. The primary vision animating HifiStell is the integration of state-of-the-art stellarator optimization science with engineering constraints which can lay the foundation for innovative stellarator fusion power plants (FPPs). We propose: • First-principles and high-fidelity simulations, several of which require exascale computing of high beta nonlinear MHD stability, core and edge turbulent transport, and energetic particle confinement, • Development of reduced order models (ROMS) which will enable intelligent searches and physics based-proxies in the vast and flexible parameter space of stellarator configurations, • Integration of the configurations with favorable physics properties with engineering constraints imposed by coils, divertor, and blanket, and • The development, verification, and validation of simulations against experiments. To address these challenges, we have brought together a complementary set of codes based on fluid as well as fully kinetic models, which comprise the HifiStell Toolkit. Some of these codes are already fully developed and performant on state-of-the-art high-performance computing platforms. Other codes are at an earlier stage of development (compared with their tokamak counterparts). They will need further algorithmic, structural advances, and validation (with uncertainty quantification) for which a research program is proposed. The further development of the latter group of codes will be facilitated by a partnership with applied mathematicians and computer scientists from the FastMath and RAPIDS2 SciDAC-5 Institutes who have played leading roles in the DOE Exascale Computing Project (ECP) on the high-fidelity whole device model of fusion plasmas (WDMApp), which has achieved a remarkable and proven record of performance on Frontier. Furthermore, the code-coupling and data science capabilities developed in the ECP WDMApp project will be leveraged to carry out coupling (both weak and strong) between the various codes in the context of a workflow. The project will be milestone-driven for all its research activities. Last but not least, HifiStell will emphasize diversity, equity, and inclusion (DEI) at all levels, as spelled out in a multi-institutional PIER plan. The team at the present time includes a woman institutional PI and six (self-identified) persons of color. The research program includes, at the outset, a prominent HBCU which will play a leading role in divertor physics studies. Every effort will be made to fill postdoctoral and graduate student openings with due attention to DEI.

HiFiStell: High-Fidelity Simulations of Stellarators

Amitava Bhattacharjee, Princeton University
David Bindel, Cornell University
Scott A. Klasky, Oak Ridge National Laboratory
Javier Hernandez Nicolau, University of California – Irvine
Elizabeth J. Paul, Columbia University
Alkesh Punjabi, Hampton University
Aaron Scheinberg, Jubilee Software Development, Inc.
Mark Shephard, Rensselaer Polytechnic Institute
Carl Sovinec, University of Wisconsin – Madison
Ben Zhu, Lawrence Livermore National Laboratory

In the stellarator, the confining magnetic field is produced mainly by external current-carrying coils. Since the stellarator can be designed to be current-free, it is intrinsically steady-state, has low recirculating power, and is free of current-driven disruptions, which probably presents the most serious challenge to tokamak operation. On the other hand, stellarators are intrinsically three-dimensional (3D) in that there is no rotational symmetry, which makes the engineering design challenging. Over the last few decades, a new concept has emerged in stellarator design, giving rise to a renaissance: the remarkable discovery that 3D magnetic confinement devices with quasisymmetry can have the same virtues as tokamaks while overcoming some of their inherent drawbacks. The current SciDAC project, HifiStell (High Fidelity Simulations of Stellarators), is a multi-institutional effort consisting of seven Universities, two DOE national laboratories, and a software company, focusing on some key science and engineering challenges pertaining to stellarators. The primary vision animating HifiStell is the integration of state-of-the-art stellarator optimization science with engineering constraints which can lay the foundation for innovative stellarator fusion power plants (FPPs).

We propose:
• First-principles and high-fidelity simulations, several of which require exascale computing of high beta
nonlinear MHD stability, core and edge turbulent transport, and energetic particle
confinement,
• Development of reduced order models (ROMS) which will enable intelligent searches and physics
based-proxies in the vast and flexible parameter space of stellarator configurations,
• Integration of the configurations with favorable physics properties with engineering constraints
imposed by coils, divertor, and blanket, and
• The development, verification, and validation of simulations against experiments.

To address these challenges, we have brought together a complementary set of codes based on fluid as well as fully kinetic models, which comprise the HifiStell Toolkit. Some of these codes are already fully developed and performant on state-of-the-art high-performance computing platforms. Other codes are at an earlier stage of development (compared with their tokamak counterparts). They will need further algorithmic, structural advances, and validation (with uncertainty quantification) for which a research program is proposed. The further development of the latter group of codes will be facilitated by a partnership with applied mathematicians and computer scientists from the FastMath and RAPIDS2 SciDAC-5 Institutes who have played leading roles in the DOE Exascale Computing Project (ECP) on the high-fidelity whole device model of fusion plasmas (WDMApp), which has achieved a remarkable and proven record of performance on Frontier. Furthermore, the code-coupling and data science capabilities developed in the ECP WDMApp project will be leveraged to carry out coupling (both weak and strong) between the various codes in the context of a workflow. The project will be milestone-driven for all its
research activities.

Last but not least, HifiStell will emphasize diversity, equity, and inclusion (DEI) at all levels, as spelled out in a multi-institutional PIER plan. The team at the present time includes a woman institutional PI and six (self-identified) persons of color. The research program includes, at the outset, a prominent HBCU which will play a leading role in divertor physics studies. Every effort will be made to fill postdoctoral and graduate student openings with due attention to DEI.