Study on the Diagnosis Method of Aero-engine Health Status Based on Stacking Ensemble Learning

Effective health status diagnosis of the aero-engine not only helps improve the safety and reliability of aero-engines, but also helps engineers and maintenance workers reduce engine maintenance and support costs. Firstly, this paper proposes integrating five different base learners based on the Stacking method to diagnose the health status of the aero-engine. Then, the deep neural network (DNN) is used to learn the complex nonlinear relationship between the base learners in Stacking ensemble (SE) learning. Finally, a case study shows that the established ensemble model has higher diagnostic stability, generalization ability and strong learning ability, and proves to be effective in health status diagnosis of aero-engines.