Pulse preparation workflows for SPARC Plasma Operations

ORAL

Abstract

Pulse preparation is critical for delivering efficient, robust operation of the SPARC tokamak. Desired pulse scenarios, which underpin Plasma Operation Plans (POPs), are developed in MOSAIC, a highly adaptable and collaborative framework designed for constructing rapid, full-pulse prediction workflows. Pulse preparation begins with the Pulse Planning Workflow (PPW), where optimization algorithms determine full-pulse trajectories that fulfill the pulse objective while maximizing the margin to system and plasma limits. The core components of PPW are the open-source codes GSPulse and TORAX. The generated actuator trajectories subsequently define the targets and parameters for the reduced models necessary for real-time, closed-loop control calculations within the Pulse Simulating Workflow (PSW). PSW comprises a collection of "episodes" engineered to test the controller's robustness against anticipated disturbances, thereby assessing the plasma control system's (PCS) resilience and its capacity for off-normal detection and response within the pulse scenarios. The outcomes of these simulations guide improvements to both the target scenario and the control schemes, thereby enhancing the probability of success. The prepared pulse is accompanied by a PCS configuration appropriate for operational deployment and provides information regarding budget consumption and device protection constraints relevant to operations.

*Funded by Commonwealth Fusion Systems

Presenters

  • Devon J Battaglia

    • Commonwealth Fusion Systems

Authors

  • Devon J Battaglia

    • Commonwealth Fusion Systems
  • Anna Teplukhina

    • Commonwealth Fusion Systems
  • Josiah Wai

    • Commonwealth Fusion Systems
  • Dan D Boyer

    • Commonwealth Fusion Systems
  • Misha Veldhoen

    • Commonwealth Fusion Systems
  • Panagiotis Stilianos Kaloyannis

    • Commonwealth Fusion Systems
  • Wil Koch

    • Commonwealth Fusion Systems
  • Ivana Abramovic

    • Commonwealth Fusion Systems
  • Thomas Alfred John Body

    • Commonwealth Fusion Systems
  • Tom Looby

    • Commonwealth Fusion Systems
  • Ryan M Sweeney

    • Commonwealth Fusion Systems
  • Christopher Woodall

    • Commonwealth Fusion Systems
  • Jonathan Citrin

    • Google DeepMind
  • Federico Felici

    • Google DeepMind