Machine Learning-Based Fault Diagnosis Approach for Geosynchronous Satellite Power Systems

Power systems, known as the hearth of satellites, have a direct impact on how long the satellite will operate. An on-orbit mission might fail due to several problems with the satellite power system. Thus, during the entire lifespan of the satellite, the diagnostics of power system faults is crucial. In this study, a new machine learning-based fault diagnosis approach has been proposed for geosynchronous (GEO) satellite power systems. In the feature extraction step of the approach, principal component analysis (PCA) technique is used. Then, LogitBoost with random forest classifier is utilized for the aim of classification. The experimental results show that the proposed model is effective and can be used for fault diagnosis of GEO satellite power systems.