Predictive Maintenance Strategy for Aero-Engines Considering Remaining Useful Life Prediction Interval  
Author Lubing Wang

 
Co-Author(s) Xufeng Zhao

 
Abstract Reliable and accurate aero-engine remaining useful life (RUL) prediction has a key role in the predictive maintenance of aero-engine. However, due to the complexity of the aero-engine operating environment, prediction errors are inevitable in traditional point estimation prediction methods. To improve the accuracy and credibility of RUL prediction, a new RUL prediction interval method is proposed to quantify the uncertainty in RUL prediction and integrate the RUL interval prediction results into predictive maintenance planning. First, we estimate the RUL prediction intervals using a deep learning ensemble model with Monte Carlo dropout. In the predictive maintenance strategy, the constructed RUL distribution is linked to the maintenance plan by considering the maintenance cost rate. The best maintenance strategy is determined by optimizing the maintenance cost rate for the maintenance economy requirement. Finally, the proposed method is validated using the aero-engine C-MAPSS dataset and can significantly reduce the maintenance cost of aero-engines.

 
Keywords Predictive maintenance, aero-engines, remaining useful life, deep learning, maintenance cost optimization
   
    Article #:  DSBFI25-8
 
Proceedings of 3rd ISSAT International Conference on Data Science in Business, Finance and Industry
January 6-8, 2025 - Da Nang, Vietnam