BEGIN:VCALENDAR VERSION:2.0 PRODID:-//linuxsoftware.nz//NONSGML Joyous v1.4//EN BEGIN:VEVENT SUMMARY:Reduced Models for Plasmas: Integration of "particle-in-cell" Simu lations and Machine Learning DTSTART:20251124T160000Z DTEND:20251124T180000Z DTSTAMP:20260906T085415Z UID:7009b9f0-cd85-475c-8f2b-9ca1cf0ee416 SEQUENCE:2 CREATED:20251121T152821Z DESCRIPTION:Accurate reduced models are essential to describe plasma dynam ics across scales while maintaining computational feasibility for large-sc ale simulations. These models require closure relations that connect macro scopic fluid quantities to the underlying kinetic behaviour\, and recent d ata-driven approaches have shown potential for deriving such relations dir ectly from kinetic simulations. In this work\, a sparse-regression framewo rk based on the Sparse Identification of Nonlinear Dynamics (SINDy) algori thm is adapted to infer fluid equations from fully kinetic OSIRIS Particle in-Cell simulations. The method is first benchmarked using the two-stream instability\, extending previous work by accurately reconstructing the hie rarchy of fluid moments up to the third-order equation. Using both the two -stream instability and a linear electron–plasma–wave configuration\, we assess under which conditions sparse regression recovers the hierarchy of fluid equations with a waterbag closure and identify its limits as non- linear phase mixing develops. We show that sparse regression can extract c ompact and physically interpretable reduced plasma models directly from ki netic simulations\, providing a data-driven path toward machine-learned cl osures applicable to future relativistic and hybrid plasma studies. LAST-MODIFIED:20251121T152832Z LOCATION:Sala V0.07 - Pavilhão de Civil URL:http://df.vps.tecnico.ulisboa.pt/pt/eventos/reduced-models-for-plasmas -integration-of-particle-in-cell-simulations-and-machine-learning/ X-ALT-DESC;FMTTYPE=text/html:

Accurate reduced mo dels are essential to describe plasma dynamics across scales while maintai ning computational feasibility for large-scale simulations. These models r equire closure relations that connect macroscopic fluid quantities to the underlying kinetic behaviour\, and recent data-driven approaches have show n potential for deriving such relations directly from kinetic simulations.

In this work\, a sparse-regression framework based on the Spars e Identification of Nonlinear Dynamics (SINDy) algorithm is adapted to inf er fluid equations from fully kinetic OSIRIS Particlein-Cell simulations. The method is first benchmarked using the two-stream instability\, extendi ng previous work by accurately reconstructing the hierarchy of fluid momen ts up to the third-order equation.

Using both the two-stream ins tability and a linear electron–plasma–wave configuration\, we assess u nder which conditions sparse regression recovers the hierarchy of fluid eq uations with a waterbag closure and identify its limits as non-linear phas e mixing develops. We show that sparse regression can extract compact and physically interpretable reduced plasma models directly from kinetic simul ations\, providing a data-driven path toward machine-learned closures appl icable to future relativistic and hybrid plasma studies.

END:VEVENT END:VCALENDAR