Predictive Maintenance Scheduling for Aircraft Engines Based on Remaining Useful Life Prediction
Lubing Wang, Ying Chen, Xufeng Zhao, Jiawei Xiang
Abstract
This paper presents a novel data-driven predictive maintenance scheduling framework for aircraft engines based on remaining useful life (RUL) prediction. First, a deep learning ensemble model is proposed to effectively predict aircraft engine RUL, including a one-dimensional convolutional neural network (CNN) and a bidirectional long short-term memory network with an attention mechanism (Bi-LSTM-AM). Second, we propose a Bayesian optimization method to optimize the hyperparameters in the deep learning ensemble model to further improve RUL prediction performance. As the aircraft engine RUL decreases over time and eventually triggers a maintenance alarm threshold. The maintenance scheduling task is initiated after the aircraft engine maintenance alert threshold has been triggered. To effectively implement the maintenance scheduling plan, we develop a novel and effective mixed-integer linear programming (MILP) model to cope with aircraft engine maintenance scheduling, which aims to minimize the maximum maintenance time. Finally, experimental results show that our proposed data-driven predictive maintenance scheduling framework can monitor the running status of aircraft engines in real time and reduce their maintenance time.