Interpretable health assessment method with influence factor traceability

Le résumé fourni par la source

The evidential reasoning method is a classical information fusion method, that can provide decision-makers with understandable and reliable equipment health state assessment results. However, the high-dimensional, correlation data collected in the equipment poses a challenge to the parameter setting and feature interpretability of evidential reasoning. At the same time, the complexity of the corresponding relationship between features and variables also increases, which makes the traceability of assessment results face challenges. To solve these problems, an interpretable health state factor traceability model is proposed. Firstly, an interpretable health assessment model based on sparse principal component analysis is proposed, which makes the statistical features of the evidence correspond to a few variables. Then, a health state factor traceability model based on variable contribution is proposed, and sensitivity contribution is defined to reflect the impact of variables on assessment results. Finally, the effectiveness of the proposed method is verified by the motor rotor experiment.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.