Ultraviolet observation (left) and magnetic field (right). The magnetic field lines are derived from simulations and show agreement with structures revealed in EUV (from Jarolim et. al, 2023).
Scientists at the University of Graz, in collaboration with Skoltech researchers, have achieved a major breakthrough in solar physics by using artificial intelligence to simulate the magnetic field in the Sun’s upper atmosphere in real time. This groundbreaking research, published in Nature Astronomy , holds great promise for advancing our understanding of the Sun’s behavior and influence on space weather.
The Sun’s magnetic field is a major driver of space weather, which can damage critical infrastructure such as electricity, aviation, and space-based technology. The main source of severe space weather events are solar active regions, which are strong magnetic fields around the sun’s surface. Current observational capabilities only allow us to measure the magnetic field on the Sun, but the buildup and release of energy occurs higher in the Sun’s atmosphere, the solar corona.
Leveraging the power of physics-savvy neural networks, the team successfully combined optical data with a physics-free magnetic field model to provide a broader understanding of the observed phenomena and the relationship to the underlying physics that governs the Sun’s motion. This cutting method is a major step in solar physics and opens up new possibilities for digital images of the sun.
The researchers simulated the observed evolution of the Sun’s active region and showed that they could perform energy-independent magnetic field simulations in real time. Remarkably, this process only requires less than 12 hours of computing time to simulate a five-day series of observations. This unprecedented speed allows scientists to conduct real-time analysis and forecasting of the Sun’s motion, enhancing our ability to predict space weather events.
The team studied the evolution of the free magnetic energy associated with solar flares on the Sun, such as coronal mass ejections, large clouds of plasma ejected from the Sun’s atmosphere at 100-3,500 km/s. . Comparison with ultraviolet observations confirmed the robustness and accuracy of the method. Crucially, the results showed a decrease in free magnetic energy both spatially and temporally, which is directly related to solar flares.
Magnetic field lines parallel to the surface magnetic field shown below (from Jarolim et. al, 2023).
Lead researcher Robert Jarolim said of the implications of this discovery: “The use of artificial intelligence in this context represents a breakthrough. Applying AI techniques to numerical simulations allows us to better process audience data and has great potential to further enhance our simulation capabilities. “The computing speed holds great promise for improving space weather forecasting and increasing our knowledge of solar behavior,” said Skoltech Associate Professor Tatiana Podlachikova.
The research, conducted by scientists at the University of Graz and Skoltech, marks a remarkable advance in the field of solar physics. Using the power of neural networks based on data from AI and physics, they discovered ways to simulate the solar corona’s magnetic field in real time, revolutionizing our ability to understand solar activity.
The research was carried out with the help of the Solar Physics Research Integrated Network Group (SPRING), supported by Scoltech’s High Performance Cluster ZORES, using solar tracking technology. SPRING was pursued in the SOLARNET project dedicated to the European Solar Telescope initiative, funded by the European Union’s research and innovation funding program Horizon 2020.
[ad_2]