Linking Single Event In-Orbit Data, Space Weather and Satellite Operations

Abstract

This thesis studies single events (SE) in an SRAM from the identical satellites Sentinel 2A and 2B, in order to find links between SE, space weather, satellite operation and device properties from time and position of event occurrence. A study of theory on single event rate prediction and space radiation environments lies the foundation for the subsequent data analysis. A focus is placed on aspects relevant with the Sentinel 2 satellites. Using Python libraries, raw satellite housekeeping data is pre-processed to make available time and place of event occurrence. Then, a moving-average filter is applied to reveal the long-term behavior of SE position and rate. The filtered SE rate data shows a linear trend, attributed to solar cycle and device degradation. Further, there are features attributed to memory hot spots, South Atlantic Anomaly (SAA) semi-annual behavior and geomagnetic storms. Future event rate data and data on affected memory addresses will help to further distinguish these influences. Filtered event positions track the drift of the SAA, but contain suspicious differences between the satellites. These are attributed to memory hot spots as well. Spectral analysis reveals that the only short-term influences in the unfiltered data arise from the orbits of the satellites. Within the SAA, event rate is similar for eclipsed and sunlit, but outside of the SAA, event rate is approximately twice as high in sunlit. This difference is currently unexplained.