Research on Anomaly Detection Method for Satellite Power Supply Based on Bayesian Model

Abstract

Abstract Satellite power system is an important system of satellites and it has a significant impact on safe and reliable operations of satellites. Researches on anomaly detection method of satellite power systems can improve the reliability of normal satellite operations. Fixed top and bottom limitations for telemetry parameters are often used in engineering to detect anomalies of telemetry parameters, but the detection thresholds do not change with time and can neither reflect the dynamic tendency of telemetry parameters nor anomalies that occur during dynamic changes of telemetry parameters. In order to solve this problem, this paper puts forward a dynamic detection method for finding anomalous satellite power supply telemetry parameters by comparing predicted values and real-time telemetry data by using dynamic characteristics of satellite power supply telemetry parameters and the Bayesian Model. This method can be used to verify telemetry data of voltage parameters of the storage battery of satellite actually in orbit. The results show that this method can effectively detect anomalies of satellite power supply.