BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Physics-informed neural networks for solar magnetic field simulati
 ons
DTSTART;VALUE=DATE-TIME:20230512T094500Z
DTEND;VALUE=DATE-TIME:20230512T100000Z
DTSTAMP;VALUE=DATE-TIME:20240704T115655Z
UID:indico-contribution-350@meetings.aip.de
DESCRIPTION:Speakers: Robert Jarolim ()\nPhysics-informed neural networks 
 (PINNs) provide a novel approach for numerical simulations\, tackling chal
 lenges of discretization and enabling seamless integration of noisy data a
 nd physical models (e.g.\, partial differential equations). In this presen
 tation\, we highlight the new opportunities for solar magnetic field model
 ling that are enabled through physics-informed machine learning.\nWe prese
 nt a novel approach for coronal magnetic field extrapolation using physics
  informed neural networks. The neural network is optimized to match observ
 ations of the photospheric magnetic field vector at the bottom-boundary\, 
 while simultaneously satisfying the force-free and divergence-free equatio
 ns in the entire simulation volume. We demonstrate that our method can acc
 ount for noisy data and deviates from the physical model where the force-f
 ree magnetic field assumption cannot be satisfied.\nWe utilize meta-learni
 ng concepts to simulate the evolution of the active region 11158. Our simu
 lation of 5 days of observations at full cadence\, requires about 12 hours
  of total computation time. The derived evolution of the free magnetic ene
 rgy and helicity in the active region\, shows that our model captures flar
 e signatures\, and that the depletion of free magnetic energy spatially al
 igns with the observed EUV emission. With this we present the first method
  that can perform realistic magnetic field extrapolations in quasi real-ti
 me\, which enables advanced space weather monitoring. \nWe conclude with a
 n outlook on our ongoing work where we extend this approach to create a ne
 w class of MHD simulations\, that can flexibly incorporate additional obse
 rvational constraints and perform fast computations.\n\nhttps://meetings.a
 ip.de/event/24/contributions/350/
LOCATION:Haus H\, Telegrafenberg
URL:https://meetings.aip.de/event/24/contributions/350/
END:VEVENT
END:VCALENDAR