Pulse Planning and Simulation for the First Campaign of SPARC

POSTER

Abstract

The mission of the first SPARC campaign is to commission systems and achieve Q > 1 in an L-mode scenario. Pulse scenarios for the first operational campaign of SPARC are prepared using MOSAIC, a flexible, collaborative framework for fast, full-pulse calculations. MOSAIC couples reduced models written in different languages to accomplish both Pulse Planning and Pulse Simulating calculations. Pulse Planning workflows use optimization algorithms to identify full-pulse trajectories that satisfy the mission of the pulse while maximizing margin to system limits, such as the requirements for plasma facing component protection. Pulse Planning workflows provide targets for Pulse Simulating workflows, where the SPARC plasma control system is coupled to models for complete closed-loop calculations. The models focus on the low- and medium-fidelity that enable rapid simulations on a single processor. Pulse Simulating workflows qualify that the control system can deliver the mission of the pulse and support the development and testing of disruption prediction and response algorithms. A simulation suite called POPSIM leverages the JAX language and the newly released TORAX code to produce rapid full-pulse simulations for operational scoping and sensitivity analysis within MOSAIC.

Presenters

  • Devon J Battaglia

    Commonwealth Fusion Systems

Authors

  • Devon J Battaglia

    Commonwealth Fusion Systems

  • Thomas Alfred John Body

    Commonwealth Fusion Systems

  • Dan D Boyer

    Commonwealth Fusion Systems

  • Christoph Hasse

    Commonwealth Fusion Systems

  • Panagiotis Stilianos Kaloyannis

    Commonwealth Fusion Systems

  • Tom Looby

    Commonwealth Fusion Systems

  • Ryan M Sweeney

    Commonwealth Fusion Systems

  • Anna A Teplukhina

    Commonwealth Fusion Systems

  • Josiah Wai

    Commonwealth Fusion Systems

  • Chris Woodall

    Commonwealth Fusion Systems

  • Jonathan Citrin

    Google DeepMind

  • Federico Felici

    Google DeepMind

  • Alex R Saperstein

    Massachusetts Institute of Technology

  • Allen Wang

    Massachusetts Institute of Technology