MHD simulations of (left) CME generation and (right) 3D magnetic reconnection performed within the HiFi modeling framework.  The smallest dynamically relevant scale for reconnection is likely meters and the largest CME-relevant scale is 1000's of megameters.  Both reconnection and CME-driven shocks produce non-thermal particle distribution functions that need to be modeled with kinetic simulation methods.
MHD simulations of (left) CME generation and (right) 3D magnetic reconnection performed within the HiFi modeling framework. The smallest dynamically relevant scale for reconnection is likely meters and the largest CME-relevant scale is 1000's of megameters. Both reconnection and CME-driven shocks produce non-thermal particle distribution functions that need to be modeled with kinetic simulation methods.

Objective
Develop, maintain, verify, and validate a state-of-the-art HPC numerical framework for fluid- and particle-based predictive modeling of space weather and plasma confinement devices

To do so, build a computational plasma framework designed to maximize the combined effectiveness of its three core features:

  1. a scalable, efficient, and accurate underlying numerical algorithm;
  2. an easy-to-use flexible interface for achieving basic and applied research objectives;
  3. modular structure that allows to improve and upgrade the framework in-step with the latest computer science and hardware advances

Approach

  • Build on the open-source HiFi multi-fluid modeling framework co-developed by NRL SSD and University of Washington for simulations of multi-scale dynamics in laboratory, solar, and heliospheric plasmas;
  • Design and implement high-order implicit and scalable numerical methods for solving fluid- and kinetic-based sets of partial differential equations. Use software libraries for grid adaptation and solving associated large sparse linear systems on the latest High Performance Computing (HPC) hardware;
  • Continue to pursue collaborations with computational physics and applied math communities for algorithm development, and with laboratory experiments for model validation

Deliverable/Value/Accomplishment

  • To achieve predictive capability with fast turn-around, multi-scale simulations of space weather and plasma confinement devices have to make full use of the O(105) core concurrency of the modern HPC hardware with scalable, adaptive algorithms in a validated modeling framework;
  • The current fluid-based version of the HiFi framework is utilized at several research laboratories and universities in the US and around the world for simulations of solar, fusion, and basic plasma phenomena. For select applications, HiFi computations have been performed on up to 104 HPC cores