Fusion reasoning algorithm for airborne correlated fault diagnosis based on meta-model

Fault diagnosis is a critical component of the maintenance and support system for complex airborne systems. Traditional diagnostic methods relying on fault transmission relationships have limitations such as monitoring blind spots and insufficient reasoning for cross-linking faults. This paper proposes a meta-model-based fusion reasoning algorithm for airborne correlated fault diagnosis, which captures system topological features via graph convolutional neural network,depictsthe fault evolution mechanisms through the meta-model, and and integrates structure features-explicit diagnostic rules-neural network reasoning by embedding association rules. Comparative tests on a typical airborne communication system show that the framework significantly improves fault isolation accuracy with strong adaptability. The research provides a new technical approach for airborne complex system fault diagnosis and supports the development of operation and maintenance technologies for similar equipment.