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:20260919T095100Z UID:7009b9f0-cd85-475c-8f2b-9ca1cf0ee416 SEQUENCE:1 CREATED:20251121T152807Z DESCRIPTION: Accurate reduced models are essential to describe plasma dyna mics across scales while maintaining computational feasibility for large-s cale simulations. These models require closure relations that connect macr oscopic fluid quantities to the underlying kinetic behaviour\, and recent data-driven approaches have shown potential for deriving such relations di rectly from kinetic simulations. In this work\, a sparse-regression framew ork based on the Sparse Identification of Nonlinear Dynamics (SINDy) algor ithm is adapted to infer fluid equations from fully kinetic OSIRIS Particl ein-Cell simulations. The method is first benchmarked using the two-stream instability\, extending previous work by accurately reconstructing the hi erarchy of fluid moments up to the third-order equation. Using both the tw o-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 compact and physically interpretable reduced plasma models directly from k inetic simulations\, providing a data-driven path toward machine-learned c losures applicable to future relativistic and hybrid plasma studies. LAST-MODIFIED:20251121T152807Z LOCATION:Sala V0.07 - Pavilhão de Civil URL:http://df.vps.tecnico.ulisboa.pt/en/events/reduced-models-for-plasmas- integration-of-particle-in-cell-simulations-and-machine-learning/ X-ALT-DESC;FMTTYPE=text/html:

Accurate reduced m odels are essential to describe plasma dynamics across scales while mainta ining computational feasibility for large-scale simulations. These models require closure relations that connect macroscopic fluid quantities to the underlying kinetic behaviour\, and recent data-driven approaches have sho wn potential for deriving such relations directly from kinetic simulations .

In this work\, a sparse-regression framework based on the Spar se Identification of Nonlinear Dynamics (SINDy) algorithm is adapted to in fer fluid equations from fully kinetic OSIRIS Particlein-Cell simulations. The method is first benchmarked using the two-stream instability\, extend ing previous work by accurately reconstructing the hierarchy of fluid mome nts up to the third-order equation.

Using both the two-stream in stability and a linear electron–plasma–wave configuration\, we assess under which conditions sparse regression recovers the hierarchy of fluid e quations with a waterbag closure and identify its limits as non-linear pha se mixing develops. We show that sparse regression can extract compact and physically interpretable reduced plasma models directly from kinetic simu lations\, providing a data-driven path toward machine-learned closures app licable to future relativistic and hybrid plasma studies.

END:VEVENT END:VCALENDAR