Adaptive and Compressed Kinetic Simulations
Abstract
This presentation overviews recent progress and future directions at the Air Force Research Laboratory's In-Space Propulsion Branch in developing accelerated, compressed models of rarefied fluid and plasma flows. The talk first describes recent work by Taitano while working at Los Alamos National Laboratory for development of adaptive moving mesh discretization of the Vlasov-Fokker-Planck Equation to be adapted to thruster applications. The talk then describes progress in accelerating solution of the Boltzmann collisional integral through the construction of stabilized Galerkin reduced order models (ROMs) and Neural Network approximations of the collision operator. Finally, streaming nonlinear state-based spatiotemporal variance reduction and the further extension of these concepts towards the construction of transient data-driven compression of kinetic flow information extracted from noisy particle simulations is described.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 25, 2021
- Accession Number
- AD1174387
Entities
People
- Alex Alekseenko
- Robert R. S. Martin
- William Taitano
Organizations
- Air Force Research Laboratory