An Anomaly Detection Algorithm of Satellite Power System Based on CUSUM Control Chart

Abstract

According to the characteristics of the power supply subsystem, this paper proposes a CUSUM control chart approach to tackle the anomaly detection and fault early warning for the satellite power supply subsystem. Firstly, through the comparison and analysis about satellites remote sensing data, this paper selects power hydrogen pressure value as a characteristic. Secondly, this paper establishes the CUSUM control chart of the power hydrogen pressure, and then detects anomaly in the satellite power system using this CUSUM control chart. Analysis of satellite telemetry data shows that when the satellite state is abnormal, the value of the characteristic will be changed and this can be recognized by the method presented in this paper, as well as the fault early warning will be given. That achieves the purposes on the satellite power system anomaly detection and fault early warning.