Aero-engine Health Monitoring Method Based on E-Bayes and DNN Fusion Decision

As a complex system, aero-engine running condition affects flight safety, so aero-engine health monitoring is necessary. We proposed an aero-engine health monitoring method based on E-Bayes method and deep neural network (DNN). Firstly, based on the fleet operation records, E-Bayes was used to calculate the reliability of aero-engine operation. Secondly, we constructed the DNN network base on the parameters collected by sensors and reliability parameter, fused the DNN results under different parameters according to the characteristics of different health condition of the aero-engine, and finally the fusion decision model based on the E-Bayes method and DNN was realized. We trained and verified the network with 9616 aero-engine running data samples contaminated by noise. The average accuracy was 96.15%, which shows that this method has good robustness.