< Back to previous page


HELIOSKILL: Heliophysics simulations and artificial intelligence

One of the most important open questions in astrophysics is explaining why the solar wind is so hot. The solar surface is at about 6500K but quickly as one moves in the atmosphere of the Sun the temperature reaches millions of degrees and continues to increase more gently at larger distances before slowly starting to decrease. This goes against any intuition for an expanding gas that moves away from its source of heating. Two new missions have been launched to understand the physics behind this puzzle more in depth: Solar Orbiter (led by Europe) and Parker Solar Probe (led by the USA). These missions are going closer to the sun than any man made object before and will give us an unprecedented wealth of data. The objective of our project is to use the mission data, supplemented by our high-resolution simulations, to extricate the holy grail: what is the equation of state of the expanding solar wind? This equation will tell us how the temperature of the solar wind evolves given the other parameters measured. The novelty of our approach is that we derive this equation with a new machine learning tool that extracts physical laws from data. Potentially this approach can bring new discovery in the form of an equation that we can then interpret and use to explain the thermodynamics of the solar wind.

Date:1 Jan 2023 →  Today
Keywords:Equation of state, machine learning, data mining, expansion of the solar wind, physics based machine learning
Disciplines:Computational physics, Data mining, Machine learning and decision making, Physics of gases, plasmas and electric discharges not elsewhere classified, Space plasma physics and solar physics