Using artificial intelligence, researchers at the University of Graz and Skoltech were able to simulate the magnetic field in the sun’s upper atmosphere in quasi-real time, resulting in a significant advance in solar physics. The results of this study, which were published in Nature Astronomy, hold a lot of promise for improving our comprehension of how the sun behaves and how it affects space weather.
The sun-powered attractive field is the primary driver of room climate, which can harm basic foundations like power, flying, and our space-based innovation. The primary wellspring of extreme space climate occasions are sun-oriented dynamic districts, which are locales around sunspots where solid attractive fields arise through the sun-powered surface. We are only able to measure the magnetic field at the surface of the sun with the tools we currently have, but the sun’s corona, or solar atmosphere, is where the energy is created and released.
The team was able to successfully integrate observational data with the physical force-free magnetic field model by making use of the capabilities of physics-informed neural networks. This resulted in a comprehensive comprehension of the connection that exists between the observed phenomena and the fundamental physics that governs the activity of the sun. This state-of-the art technique denotes a critical achievement in sun-oriented physical science and opens up new doors for mathematical reenactments of the sun.
“Our application of artificial intelligence in this context represents a quantum leap forward. The application of AI approaches to numerical simulations helps us to better incorporate observational data and has the potential to significantly increase our simulation skills.”
Robert Jarolim, the lead researcher,
The researchers demonstrated the capability of real-time force-free magnetic field simulations by simulating the evolution of an observed solar active region. Stunningly, this interaction just required under 12 hours of calculation time to recreate a perception series of five days. Our ability to anticipate space weather events is enhanced by scientists’ ability to analyze and forecast solar activity in real time, thanks to this unprecedented speed.
Observed surface magnetic field at the bottom of the simulated magnetic field lines. Credit: Jarolim et al., 2023
The group additionally concentrated on the time advancement of free attractive energy inside the coronal volume, which is connected to sun-powered eruptive occasions on the sun like coronal mass discharges—huge plasma mists launched out from the air of the sun at rates of 100–3,500 km/s. The correlation with outrageously bright perceptions affirmed the power and exactness of the system. Importantly, significant spatial and temporal depletions of free magnetic energy were found to be correlated with observed solar eruptions.
Robert Jarolim, the lead specialist, said, “Our utilization of computerized reasoning in this setting addresses an extraordinary leap forward. The utilization of man-made intelligence strategies for mathematical recreations permits us to more readily consolidate observational information and holds incredible potential for additional development of our reproduction capacities.”
Skoltech Academic Partner Tatiana Podlachikova says, “The figuring speed holds critical commitment for further developing space weather conditions, determining and propelling our insight into the sun’s way of behaving.”
The findings of this study, which were carried out by researchers from Skoltech and the University of Graz, represent a significant advancement in solar physics. They have created real-time simulations of the solar coronal magnetic field by utilizing the power of AI and physics-informed neural networks. This has revolutionized our comprehension of solar activity.
More information: R. Jarolim et al, Probing the solar coronal magnetic field with physics-informed neural networks, Nature Astronomy (2023). DOI: 10.1038/s41550-023-02030-9
