We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

physics.plasm-ph

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Physics > Plasma Physics

Title: Identification of high order closure terms from fully kinetic simulations using machine learning

Abstract: Simulations of large-scale plasma systems are typically based on a fluid approximation approach. These models construct a moment-based system of equations that approximate the particle-based physics as a fluid, but as a result lack the small-scale physical processes available to fully kinetic models. Traditionally, empirical closure relations are used to close the moment-based system of equations, which typically approximate the pressure tensor or heat flux. The more accurate the closure relation, the stronger the simulation approaches kinetic-based results. In this paper, new closure terms are constructed using machine learning techniques. Two different machine learning models, a multi-layer perceptron and a gradient boosting regressor, synthesize a local closure relation for the pressure tensor and heat flux vector from fully kinetic simulations of a 2D magnetic reconnection problem. The models are compared to an existing closure relation for the pressure tensor, and the applicability of the models is discussed. The initial results show that the models can capture the diagonal components of the pressure tensor accurately, and show promising results for the heat flux, opening the way for new experiments in multi-scale modeling. We find that the sampling of the points used to train both models play a capital role in their accuracy.
Comments: 37 pages, 12 figures, 2 tables
Subjects: Plasma Physics (physics.plasm-ph); Machine Learning (cs.LG); Computational Physics (physics.comp-ph)
Journal reference: Physics of Plasmas 29, 032706 (2022)
DOI: 10.1063/5.0066397
Cite as: arXiv:2110.09916 [physics.plasm-ph]
  (or arXiv:2110.09916v2 [physics.plasm-ph] for this version)

Submission history

From: Brecht Laperre [view email]
[v1] Tue, 19 Oct 2021 12:27:02 GMT (1774kb)
[v2] Thu, 24 Mar 2022 11:14:21 GMT (1841kb,D)

Link back to: arXiv, form interface, contact.