Spacecraft flywheel system health assessment method based on the interpretable belief rule base with result traceability

Abstract It is crucial to accurately and reliably assess the health status of the flywheel system to ensure the safety of spacecraft in orbit. In response to the uncertainty modeling problem in the health assessment of flywheel systems, this paper introduces a rule-based modeling technique, the belief rule base (BRB), to effectively characterize its causal relationships. Considering the strict reliability requirements of aerospace systems, health assessment models should not only have high process interpretability, but also traceability of assessment results. Therefore, a spacecraft flywheel system health assessment method based on the interpretable BRB with result traceability is proposed in this article. Firstly, seven interpretability criteria were proposed based on the characteristics of the flywheel system. Secondly, three interpretability constraint strategies were designed to limit the interruption of model interpretability during the optimization process. Furthermore, to achieve reverse tracing of evaluation results, a Gaussian membership function is introduced to enhance the transparency and traceability of the model. Finally, a health status assessment experiment was conducted using the bearing components of a certain flywheel system as a case study to verify the effectiveness of the proposed method.