Abnormal Discharge Detection Method of Satellite Battery System Based on FPgrowth-RF

In order to deal with the energy imbalance caused by abnormal discharge of the battery system during in-orbit operation and improve the reliability of the battery system and even the whole satellite, it is necessary to find out the abnormal discharge state of the battery in time and accurately. Aiming at the problems of battery system data imbalance, complex relationship among data and single known mechanism, an abnormal discharge detection method of satellite battery system based on FPgrowth-RF is proposed. Firstly, the characteristics related to abnormal discharge are extracted according to the working mechanism relationship between related currents, and then the RF classification model is established after the samples are equalized. Then, based on FPgrowth algorithm, the co-occurrence rules of discharge events of different batteries are mined, and the abnormal discharge state of batteries is detected by combining the mechanism and data rules. This method can make the false detection rate less than 0.0003 and the missed detection rate less than 0.0001, and improve the detection efficiency and accuracy.