Understand and Forecast Killer Electrons- Coupling Machine Learning and First-principle Simulations
Abstract
The Earth s outer radiation belt consists of energetic electrons, also called killer electrons, that can damage or kill satellites. These killer electrons in the Van Allen radiation belts travel around at nearly light speed, and these energetic particles pack a lot of puch. Therefore, a significant increase in the number of them during geomagnetic storms could seriously harm spacecraft systems and astronauts. The main objective of this project is to study how, where, and why the energetic electrons in the Van Allen radiation belt increase during geomagnetic storms. In science language, we aim to answer an overarching scientific question- what are the key physical processes that control the dynamics of the energetic electrons in the Earth s inner magnetosphere. Although there are a lot of past studies on this topic, they are limited by the number of satellites in space. These satellites can provide a few observation points in space, but the responsible physical mechanisms are all over space, where there are not always satellites flying. Thus, the observations have limited our understanding of the radiation belt and the capability to forecast killer electrons. This project takes advantage of an artificial intelegence technique called neural networks. As already shown in our latest study, the neural network can learn how the energetic electrons respond to external drivers. Most importantly, the neural network can provide the capability to fly virtual satellites all over space, giving virtual observations of energetic electrons in space. This cannot be done in the past. This capability will push our understanding of the Van Allen radiation belt forward by comparing the virtual observations with real observations, theoretical expectations, and numerical simulations. More importantly, the neural network can provide a reliable and accurate forecast of the killer electrons and help us protect our spacecraft systems and astronauts.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 06, 2024
- Source ID
- FA95502310359
Entities
People
- Xiangning Chu
Organizations
- Air Force Office of Scientific Research
- Regents of the University of Colorado
- United States Air Force