Presentation Information

2:15 PM - 2:30 PM JST(5:15 AM - 5:30 AM UTC)Candidate for Academic Incentive Award

[B-2-15]A Comparative Study of Models for Detecting Surface Charging in Low Earth Orbit Satellites

◎Iroha Fujiwara¹, Mariko Teramoto¹, Teppei Okumura², Akitoshi Hanazawa¹ (1. Kyutech, 2. JAXA)

Keywords:

Satellite,Surface Charging,Machine Learning

Surface charging-induced discharges in low Earth orbit (LEO) satellites can lead to satellite failures. Therefore, the detection and prediction of charging events are essential. Conventionally, event extraction has been performed using a threshold method applied to observation data from electrostatic analyzers. However, this approach may fail to detect events that fall outside the threshold. In this study, using 2010 observation data from the DMSP-F16 satellite, we compared a threshold method with a CNN-based charge event extraction method that focuses on the ion-line in the Energy-Time (ET) diagrams. As a result, the threshold method extracted 433 charge events, while the ion-line method extracted 196. 119 events were detected by both methods. In the future, we aim to develop a more accurate method for detecting surface charge events by conducting a detailed analysis of the characteristics of the events extracted by each method.

Back to Session information