Fault Detection of Aeronautical Electromagnetic Relay Feature Fusion based on Evidence Reasoning

Aeronautical electromagnetic relay (AMER) is critical component in aircraft systems, whose failure can lead to severe consequences, including flight accidents. To enhance the accuracy of fault detection before takeoff, a novel method leveraging evidence reasoning has been proposed. This method establishes an evidence description framework that employs interval evidence structure, transitioning from single-value data to interval data representation. By utilizing expert knowledge, it transforms data into interval evidence using triangular fuzzy membership functions. The method then determines the fusion weight and reliability of the interval evidence, conducting multi-feature fusion through interval evidence reasoning (ER) rules. To mitigate evidence conflict, a discount factor is introduced to adjust the fusion weight, and fault detection results are derived by iteratively combining historical and current fault evidence. This interval evidence approach more effectively captures the uncertainty inherent in data compared to traditional single-valued data fusion methods. The multi-feature fusion strategy provides a comprehensive assessment of the equipment’s operational status. Experimental results have demonstrated that this method achieves high accuracy and can be used for fault detection of AMER.