Research on Satellite Power Anomaly Detection Method Based on LSTM

Due to the important role of satellite power system in satellite operations, testing the on-orbit operating status of satellite power systems will have important guiding significance for the safe and stable operation of the entire satellite and the adjustment of control strategies. Starting from the telemetry data of satellite power system, this paper proposes a method for detecting anomaly of satellite power system parameters based on Long Short Term Memory (LSTM). The LSTM prediction model is used to predict the parameters of the satellite power system, given the error between the predicted parameter data and the actual parameter data is the minimum target, the model is trained to detect the anomaly of the satellite power system, and is verified through the power system parameters’ telemetry data in a certain actual on-orbit satellite. The result shows that the method can effectively detect anomaly in satellite power system, and has high prediction accuracy and wide applicability.