Speaker
Description
The structure and dynamics of the solar atmosphere are a manifestation of the magnetic field, and it is from this field that all phenomena observed in this region are guided. Although its direct measurement in the solar corona faces limitations, its configuration can be calculated from data from the photosphere. This study proposes the use of Physics-Informed Neural Networks (PINNs) to model the coronal magnetic field.
This methodology combines the efficiency of machine learning with the differential equations of solar magnetohydrodynamics. The application of PINNs allows physical constraints to be incorporated directly into the learning process, enabling more robust inferences from observational data. The central objective of this project is to develop and validate a new PINN-based method for three-dimensional extrapolation of the coronal magnetic field. This new model provides a more accurate analysis of the interaction between the field and the atmospheric plasma, as well as to develop tools for studying its temporal evolution. This approach seeks a more precise analysis of the magnetic structure and its behavior.
