| Name: | Jithu J Athalathil |
| Affiliation: | Indian Institute of Technology Indore |
| Conference ID : | ASI2024_295 |
| Title : | Surface Flux Transport Modelling Using Physics Informed Neural Networks |
| Authors : | Jithu J Athalathil1, Bhargav Vaidya1, Sayan Kundu1 |
| Authors Affiliation: | 1 Indian Institute of Technology, Indore-453552,India |
| Mode of Presentation: | Oral |
| Abstract Category : | Sun, Solar System, Exoplanets, and Astrobiology |
| Abstract : | Studying the solar surface and its dynamic behaviour plays a crucial role in understanding the origin of space weather drivers that can have a potential impact on Earth. The initiation of these drivers is closely related to the magnetic activity of our Sun. Surface Flux Transport(SFT) modelling allows us to simulate and analyse the transport and evolution of magnetic flux on the solar surface, providing valuable insights into the mechanisms responsible for solar activity. Further, magnetic field modelling aids in constraining the solar dynamo models as well as serving as an initial condition to extrapolate the solar magnetic field to the heliosphere. These simulations also help in forecasting the magnetic activity of the upcoming solar cycle. The main physical ingredients are typically advection, diffusion, and magnetic flux emergence, using empirical relations to mimic the solar magnetic field's evolution.
We have developed a novel Physics Informed Neural Networks(PINNs)-based model to study the evolution of Bipolar Magnetic regions (BMRs) using SFT in azimuthally averaged (1D) and 2D magnetic fields. The results have been compared with standard and higher order finite difference SFT code developed using advection and diffusion algorithms. PINNs produce mesh-independent solutions with comparable accuracy to second-order numerical techniques and is computationally less intensive. We have developed a numerical model for second-order accuracy in time and space using the RK-IMEX scheme to get results comparable with PINNs. This showcases the applicability of PINNs in solving advection-diffusion equations with a particular focus on heliophysics. In this presentation, I shall present a comparison between the results obtained from the numerical approach and those from PINNs. We intend to employ the SFT module to provide initial magnetograms essential for input boundary conditions of the SWASTi framework to forecast the solar wind plasma parameters at the L1 point. |
