Neural network based prediction of solar array degradation during all-electric satellite deployment

Loading...
Thumbnail Image
Issue Date
2020-05-01
Authors
Farabi, Tanzimul Hasan
Advisor
Dutta, Atri
Citation

Farabi, T. 2020. Neural network based prediction of solar array degradation during all-electric satellite deployment -- In Proceedings: 16th Annual Symposium on Graduate Research and Scholarly Projects. Wichita, KS: Wichita State University, p.25

Abstract

In recent years solar electric propulsion (SEP) has become a topic of great interest among researchers in the satellite and space industry due to its advantages over traditional chemical propulsion system. All electric satellites allow for designing light weight and small sized orbiters that can significantly reduce launch costs. Due to its superior propellant management SEP has now become an option for low thrust orbit raising mission. However the key issue with all electric satellite is its long transfer time spanning several months. Moreover this long transit exposes the electrical components and solar arrays of the satellite to the highly energized protons and electrons trapped in the Van Allen radiation belts. The charged particles damage the solar cells, resulting in the degradation of propulsive power, thus it has become very important to characterize the power loss due to radiation damage. In this study, a radiation database for the trapped particles is created using the combination of Ap9/Ae9 radiation tool and the European Space Agency (ESA) developed SPENVIS software. The generated database is then used for training of an artificial neural network algorithm, which predicts the loss of thrust output during orbit raising maneuvers. Finally, the power degradation output from the neural network is incorporated within a low thrust trajectory optimization solver to generate optimal trajectory while estimating the updated thrust output. The findings of this study is compared with similar orbit transfer scenarios to demonstrate the impact of the work. As neural network based modeling allows for fast radiation damage computations, it will enable mission designers to conduct rapid exploration of design space, thus help in taking critical mission planning decisions.

Table of Content
Description
Presented to the 16th Annual Symposium on Graduate Research and Scholarly Projects (GRASP) held online, Wichita State University, May 1, 2020.
Research completed in the Department of Aerospace Engineering, College of Engineering
publication.page.dc.relation.uri
DOI