Aerospace Relay Health Status Assessment Method Based on SMOTE-Enhanced EBRB

The aerospace relay, a complex and precise power transmission system integrating electromagnetic and mechanical elements, holds vital significance for spacecraft safety, rendering its operational condition crucial. Therefore, it is imperative to conduct assessments of the health status of aerospace relays. However, existing models for aerospace relays are beset with issues related to data uncertainty and lack of interpretability of evaluation outcomes. Existing models for aerospace relays grapple with issues surrounding data uncertainty and the uninterpretable nature of their evaluation results. The extended belief rule base (EBRB) is a modeling approach adept at addressing uncertainties, characterized by a lucid inference mechanism and inherently interpretable results. However, current research has overlooked the fact that, during actual operation of aerospace relays, external disturbances and internal influences can lead to an imbalance among health status categories, resulting in a tendency for minority samples to be disregarded during training and rendering accurate evaluations of all categories unattainable. Consequently, this paper proposes an EBRB-based method for assessing the health status of aerospace relays that incorporates Synthetic Minority Over-sampling Technique (SMOTE). Firstly, the sample dataset is balanced using SMOTE. Secondly, an EBRB model is constructed to process the data. Lastly, ensemble reasoning (ER) algorithm is employed for fusion inference. In the experiment, the data processed by this model demonstrated obvious advantages across all performance indicators, indicating the model's powerful efficacy in the health status assessment of aerospace relays.